Google filters for golf course results are terrible so I built a better way to browse courses using OSM as the backbone.... https://golfcoursebrowser.com/
It's a work in progress and mostly US for now, but I want to expand to the rest of North America, the EU, and the rest of the world ASAP.
It's free, no ads, no login, no bs. If you spot anything wrong (bad info, a missing course, wrong scorecard), you can flag it right on the course page. I actually read those and fix them.
Still a lot of missing info, but I'm filling out more and more each day. The goal is the most complete, and current golf course directory in the world, verified and maintained by actual golfers. I think is attainable in the next few months.\nThe base layer is OpenStreetMap, cross-checked against course websites and enriched with scorecards, USGA rating/slope, and public/private status. User corrections from golfers who know their home course have beaten every commercial data source we check against.", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49053708", "k": "Alien World Chemistry Found Inside Meteorite That Struck New Jersey Home", "domain": "technology", "knowledge": "Alien World Chemistry Found Inside Meteorite That Struck New Jersey Home", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49057985", "k": "Show HN: Managing on-premise servers without Kubernetes", "domain": "technology", "knowledge": "Show HN: Managing on-premise servers without Kubernetes", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49052564", "k": "Cloudflare's new AI traffic options for customers", "domain": "technology", "knowledge": "Cloudflare's new AI traffic options for customers", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49057972", "k": "Ask HN: What are the most promising RL fields for a new master student?", "domain": "technology", "knowledge": "i am an incoming msc student thinking about a research direction in reinforcement learning. which rl subfields do you think have the most potential right now? i am particularly interested in emboddied ai and bcis.", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49034777", "k": "Stinkpot: SQLite-backed shell history", "domain": "technology", "knowledge": "Stinkpot: SQLite-backed shell history", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49050512", "k": "Running a 28.9M parameter LLM on an $8 microcontroller", "domain": "technology", "knowledge": "Running a 28.9M parameter LLM on an $8 microcontroller", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49002762", "k": "DskDitto: Ultra-fast, parallel duplicate-file detector", "domain": "technology", "knowledge": "DskDitto: Ultra-fast, parallel duplicate-file detector", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49023243", "k": "I wouldn't say Pangram is broken, but I would say that it's brittle", "domain": "technology", "knowledge": "Response: https://substack.com/@maxspero/note/c-297953357", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49013583", "k": "Capra, Cribl's Design System", "domain": "technology", "knowledge": "Capra, Cribl's Design System", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49024980", "k": "Rethinking Legal Education in the AI Era", "domain": "technology", "knowledge": "Rethinking Legal Education in the AI Era", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49056620", "k": "Terence Tao: Mathematics in the Age of AI [pdf]", "domain": "technology", "knowledge": "Terence Tao: Mathematics in the Age of AI [pdf]", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49050859", "k": "LLM Usage in Debian: Three Proposals", "domain": "technology", "knowledge": "LLM Usage in Debian: Three Proposals", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49053385", "k": "Git rebase -I is not that scary", "domain": "technology", "knowledge": "Git rebase -I is not that scary", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49028169", "k": "Pip install Postgres – no Docker/Brew/apt", "domain": "technology", "knowledge": "Pip install Postgres – no Docker/Brew/apt", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49056917", "k": "Pizza, pasta, potatoes, protein – how Italian children became so overweight", "domain": "technology", "knowledge": "Pizza, pasta, potatoes, protein – how Italian children became so overweight", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49052628", "k": "Clinical failure rates over the decades: yikes", "domain": "technology", "knowledge": "Clinical failure rates over the decades: yikes", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49005197", "k": "Overloaded Overloading", "domain": "technology", "knowledge": "Overloaded Overloading", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49013464", "k": "SIMD for Collision", "domain": "technology", "knowledge": "SIMD for Collision", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49045159", "k": "Android may soon restrict on-device ADB", "domain": "technology", "knowledge": "Android may soon restrict on-device ADB", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49053382", "k": "Systems and Delays", "domain": "technology", "knowledge": "Systems and Delays", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49049693", "k": "Show HN: Brolly, a plain-text weather forecast site", "domain": "technology", "knowledge": "The UK MET office recently redesigned their site, adding a lot of additional whitespace, scrolling, and animations. This significantly reduced its usability for me, and left me wanting an ‘at a glance’ weather site.
I made https://brolly.sh, a minimalist, plain text weather forecasting site. You can use it to view weather from around the world, with: 7 day forecast; Previous day log (so you can confirm it definitely was cooler / hotter / wetter / drier yesterday!); Hourly rain, wind, temperature, conditions; Hourly UV, air quality and pollen, including pollen type specific forecasts within the EU / UK; Location search and last 5 locations; Location specific units.
I mostly made the site for myself, if anyone else also benefits from it that’s an added advantage.
You can check out the weather in York, UK at https://brolly.sh/forecast/RWFP2qW8, or search for a location at https://brolly.sh
The site is deliberately styled as a single long scrollable column, to work on mobile phones. You can view it on desktop too, there's just a lot of horizontal padding. I naturally took a lot of inspiration from plaintextsports.com. Despite not being a sports fan, I love its aesthetic. But, you'll hopefully see that this site isn't a rip off, and has its own deliberate look and feel.
Visualisations are really important to showing information at a glance. I spent a lot of time designing the different visualisations and making them work with only characters. My favourite is the hourly heat map used for pollen count.
It's also frustrating to have an interactive site, where you can't share a page with a friend and have them see what you're seeing. To solve this, all page state (i.e. location, selected day, expanded / collapsed sections), is stored in the URL. You can share or bookmark the specific view, and know that you'll always be able to come back to it.
The site uses PocketBase. It’s written in Go and plain HTML/JavaScript/CSS. All pages are backend rendered, with light JavaScript to handle re-loading content without page jumps when using interactive features like next / previous day navigation. Weather forecasts are fetched from open-meteo.com, which has a very generous free tier. However, I also built a custom LRU cache on top of PocketBase’s SQLite DB to cache forecasts for 5 minutes, and avoid putting unnecessary pressure on the open-meteo API.", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49039868", "k": "Show HN: I made some transistor animations", "domain": "technology", "knowledge": "Hi HN,\nI made some animations of the most important kinds of transistors using my semiconductor simulation, details of which are on the page. I tried to make the visuals as realistic as possible while also aiming for clarity. If you want to go beyond the charge carriers and look at, for example, the electric field, you can do so in the simulation software. The desktop software also has less common devices like IBGTs and SCRs that have similar animations.
The last thread about my software was posted here about a year ago: https://news.ycombinator.com/item?id=43942279", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49054224", "k": "JetZero", "domain": "technology", "knowledge": "JetZero", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49051947", "k": "GM Backs Sodium Ion Batteries for U.S. Grid Storage", "domain": "technology", "knowledge": "GM Backs Sodium Ion Batteries for U.S. Grid Storage", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49050058", "k": "Memory safety absolutists", "domain": "technology", "knowledge": "Memory safety absolutists", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49006096", "k": "How My Images Are Dithered", "domain": "technology", "knowledge": "How My Images Are Dithered", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49046858", "k": "Kyber (YC W23) Is Hiring a Head of Engineering", "domain": "technology", "knowledge": "Kyber (YC W23) Is Hiring a Head of Engineering", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49047365", "k": "Bitchat is now on Radicle", "domain": "technology", "knowledge": "Bitchat is now on Radicle", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49057298", "k": "Show HN: QueryTuner – SQL diagnostics across 5 DBs, no database connection", "domain": "technology", "knowledge": "Show HN: QueryTuner – SQL diagnostics across 5 DBs, no database connection", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49058547", "k": "Hallmark – Anti-AI-Slop Design Skill for Claude Code, Cursor, and Codex", "domain": "technology", "knowledge": "Hallmark – Anti-AI-Slop Design Skill for Claude Code, Cursor, and Codex", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49048034", "k": "Open-weight AI is having its Kubernetes moment", "domain": "technology", "knowledge": "Open-weight AI is having its Kubernetes moment", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49057200", "k": "You don't hire juniors to do menial chores", "domain": "technology", "knowledge": "You don't hire juniors to do menial chores", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49058917", "k": "Repairing a \"for parts\" Canon 7D Mark II", "domain": "technology", "knowledge": "Repairing a \"for parts\" Canon 7D Mark II", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49051369", "k": "Turn And Face The Strange", "domain": "technology", "knowledge": "Turn And Face The Strange", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49051120", "k": "Did they ghost you?", "domain": "technology", "knowledge": "Did they ghost you?", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49050735", "k": "Producing ammonia and fertiliser using wind power in Morris, Minnesota", "domain": "technology", "knowledge": "Producing ammonia and fertiliser using wind power in Morris, Minnesota", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49008512", "k": "An old patent inspired the new \"Y-zipper\", a three-sided fastener", "domain": "technology", "knowledge": "https://www.youtube.com/watch?v=AWig98GVIno
https://www.tomshardware.com/3d-printing/mit-researchers-rev...
https://www.yankodesign.com/2026/05/31/a-zipper-patent-sat-i...", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49056854", "k": "Show HN: Wmux – A workspace multiplexer for AI agents", "domain": "technology", "knowledge": "Show HN: Wmux – A workspace multiplexer for AI agents", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49054697", "k": "W4ME Station – a WASM-4 runtime for Java ME phones", "domain": "technology", "knowledge": "W4ME Station – a WASM-4 runtime for Java ME phones", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49051463", "k": "Multicast TV Distribution on My Home Network", "domain": "technology", "knowledge": "Multicast TV Distribution on My Home Network", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49046525", "k": "The Fedora 45 Sausage Factory", "domain": "technology", "knowledge": "The Fedora 45 Sausage Factory", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49056340", "k": "France and Spain wildfires: more than 300k people displaced", "domain": "technology", "knowledge": "France and Spain wildfires: more than 300k people displaced", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49045446", "k": "MouthPad: A Tongue-Controlled Touchpad", "domain": "technology", "knowledge": "MouthPad: A Tongue-Controlled Touchpad", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49056664", "k": "KGazer: Kafka Compacted Topic Explorer", "domain": "technology", "knowledge": "KGazer: Kafka Compacted Topic Explorer", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49048689", "k": "Bringing PyTorch Monarch to AMD GPUs", "domain": "technology", "knowledge": "Bringing PyTorch Monarch to AMD GPUs", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49051707", "k": "Becoming a Research Engineer at a Big LLM Lab", "domain": "technology", "knowledge": "Becoming a Research Engineer at a Big LLM Lab", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49007018", "k": "Spatial languages: Writing code in 2D", "domain": "technology", "knowledge": "Spatial languages: Writing code in 2D", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49008396", "k": "Zero roadkill as Amazon canopy bridges secure 15,000 crossings", "domain": "technology", "knowledge": "Zero roadkill as Amazon canopy bridges secure 15,000 crossings", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49058593", "k": "The hacker who humiliated spyware makers and was never caught", "domain": "technology", "knowledge": "The hacker who humiliated spyware makers and was never caught", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49054908", "k": "Maybe we should revisit microkernels", "domain": "technology", "knowledge": "Maybe we should revisit microkernels", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49051130", "k": "Show HN: Writemark, a dependency free web component for inline Markdown editing", "domain": "technology", "knowledge": "I like writing Markdown, but do not like writing it inside a plain textarea.
I wanted something I could use anywhere by dropping in a single web component:
```
<writemark-editor name="body"></writemark-editor>
```
That became Writemark.
It renders Markdown while you write, but Markdown remains the value you read, store, and submit. It also has source, split, and preview modes, along with slash commands, tables, task lists, code blocks, native form support, and an API for adding your own controls. There are no runtime dependencies and no required framework or built in toolbar.
It is fully vibecoded. The process was very iterative. I knocked something out, tried using it, found bugs, fixed them, and repeated that cycle until I had something I liked writing in and that performed reasonably well.
It did not begin as an experiment about AI generated software -- just did not want to use textareas anymore. I like Markdown, and I wanted one component that I could use anywhere without bringing along an entire editor framework.
It is still very young. The parser is handwritten, the component is essentially one large JavaScript file, and I am certain there are edge cases waiting to be found. If you find one, I am happy to fix it.
I have tried to give it a decent safety net. It currently has 951 Playwright checks across Chromium, Firefox, and WebKit, along with hostile input cases, sanitizer fuzzing, and differential tests against CommonMark.
I built this because I wanted it for myself, but I think it turned into something kind of cool. I hope some of you enjoy it. I would love hearing what you think, especially if you try it and manage to break something.
- Github: https://github.com/Brostoffed/writemark
- NPM: https://www.npmjs.com/package/writemark-editor", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49014034", "k": "Charles Ross spent 50 yrs building Star Axis naked-eye observatory in New Mexico", "domain": "technology", "knowledge": "Charles Ross spent 50 yrs building Star Axis naked-eye observatory in New Mexico", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49056403", "k": "Upd.dev: Free Git hosting, including the freedom to vibe code", "domain": "technology", "knowledge": "Upd.dev: Free Git hosting, including the freedom to vibe code", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49015286", "k": "German Peasants' War", "domain": "technology", "knowledge": "German Peasants' War", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49043724", "k": "Hannah Fry Wins the Leelavati Prize in 2026 for Mathematics Outreach", "domain": "technology", "knowledge": "Hannah Fry Wins the Leelavati Prize in 2026 for Mathematics Outreach", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49054307", "k": "Humans haven't stopped evolving", "domain": "technology", "knowledge": "Humans haven't stopped evolving", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49044874", "k": "Extinct Media Museum Tokyo", "domain": "technology", "knowledge": "Extinct Media Museum Tokyo", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49056351", "k": "Systemd Linger", "domain": "technology", "knowledge": "Systemd Linger", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49056278", "k": "Apertus 1.5, Swiss open-weight, open-source, open training data model", "domain": "technology", "knowledge": "Apertus 1.5, Swiss open-weight, open-source, open training data model", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49048681", "k": "The Dark Night of Mathematics", "domain": "technology", "knowledge": "The Dark Night of Mathematics", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49049638", "k": "Show HN: SpinWin – A macOS menu bar app to visually rotate or spin any window", "domain": "technology", "knowledge": "SpinWin is an open source menu bar app to visually rotate any individual window. You can rotate / spin as many as you want one by one.
I was scrolling reels when one came on where the content creator suggested "turning the phone upside down" for a portion. I was sitting at my mac, and decided to see if I could find something that did this. To my surprise, nothing exists, so I decided to build it.
Turns out macOS doesn't provide any public API to do anything like this, so I wound up using the Accessibility API to move the target window offscreen, where it keeps rendering. Then we capture it's contents with ScreenCaptureKit, and display that on a transparent, borderless overlay window which has a CALayer to rotate it.
It's written in Swift, open source, MIT license, and signed/notarized.
Have fun - would love to hear your comments.", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49056135", "k": "Anthropic should learn from those cotton-picking socialists", "domain": "technology", "knowledge": "Anthropic should learn from those cotton-picking socialists", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49047516", "k": "My web version of Mars MIPS, now has built-in C compiler", "domain": "technology", "knowledge": "My web version of Mars MIPS, now has built-in C compiler", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49050538", "k": "The growing vigilante movement to knock out Flock surveillance cameras", "domain": "technology", "knowledge": "The growing vigilante movement to knock out Flock surveillance cameras", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49055472", "k": "Skateboarding Forum on Wood Quality", "domain": "technology", "knowledge": "Skateboarding Forum on Wood Quality", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49056036", "k": "Build a self-scaling OCR pipeline with Qwen 3.5 and Kubernetes", "domain": "technology", "knowledge": "Build a self-scaling OCR pipeline with Qwen 3.5 and Kubernetes", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49044492", "k": "UK AISI / Caisi Preliminary Assessment of Kimi K3's Cyber Capabilities", "domain": "technology", "knowledge": "UK AISI / Caisi Preliminary Assessment of Kimi K3's Cyber Capabilities", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49047563", "k": "The Silurian Hypothesis (2020)", "domain": "technology", "knowledge": "The Silurian Hypothesis (2020)", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49056022", "k": "Claude Code has a hardcoded instruction telling Opus 5 not to use subagents", "domain": "technology", "knowledge": "Claude Code has a hardcoded instruction telling Opus 5 not to use subagents", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49055927", "k": "Show HN: Reproducibility Benchmark a Risk Quantitative Model", "domain": "technology", "knowledge": "Show HN: Reproducibility Benchmark a Risk Quantitative Model", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49033004", "k": "If coding has been solved, why does software keep getting worse?", "domain": "technology", "knowledge": "If coding has been solved, why does software keep getting worse?", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49021856", "k": "I regret migrating to Codeberg", "domain": "technology", "knowledge": "I regret migrating to Codeberg", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49022152", "k": "Writing by hand is good for your brain", "domain": "technology", "knowledge": "Writing by hand is good for your brain", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49047884", "k": "Task-centered iproute2 user guide", "domain": "technology", "knowledge": "Task-centered iproute2 user guide", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49057337", "k": "Ask HN: Are you using Rust on embedded devices yet? If not, why?", "domain": "technology", "knowledge": "I started dabbling with ESP32-based MCUs from Waveshare with Rust, and I'm quite impressed with the state of things (esp-rs, embassy, probe-rs etc). Note that I haven't really done embedded in any other languages, so I don't really have anything to compare it with.
One thing I did notice was that creating beautiful, interactive user interfaces on MCUs with displays is a little harder, libs like embedded-graphics don't look that great.
This made me wonder, are people using Rust in their professional embedded work?", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49055849", "k": "Anthropic secures its AI-native software development lifecycle", "domain": "technology", "knowledge": "Anthropic secures its AI-native software development lifecycle", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49055825", "k": "Yann LeCun's Bet That Intelligence Starts in the World", "domain": "technology", "knowledge": "Yann LeCun's Bet That Intelligence Starts in the World", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49038433", "k": "Claude Opus 5", "domain": "technology", "knowledge": "https://www.anthropic.com/claude-opus-5-system-card", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49026810", "k": "Show HN: Echo – Fable-level results at 1/3 the cost using open-weight models", "domain": "technology", "knowledge": "I’ve been building Echo (https://echo.tracerml.ai/), an experiment in making one AI system out of a pool of open-weight models rather than choosing a single model and using it for every task.
It started with a simple experiment. I took a group of models, including GLM-5.2, Kimi K2.7 and others, and ran them on the same evaluations. Then I measured what would happen if, for each problem, you somehow knew in advance which models would be useful and how their outputs should be combined.
That hypothetical system performed substantially better than any individual model in the pool. Of course, it is not something you can actually deploy because it relies on knowing which decisions were good after seeing the result. Echo is my attempt to recover some of that advantage without having that information in advance.
For each request, Echo decides how much computation to allocate, which models should participate, and how their work should be combined. Some prompts may only need a relatively small amount of inference, while others benefit from multiple models working on different parts of the problem.
One thing that surprised me while building it was how complementary the models are. A model that is clearly weaker overall can still be extremely useful on particular problems or as part of a combination.
On my first evaluation mix, Echo consistently performed better than the best individual model in its pool. It also reached roughly the same aggregate result as Fable, which I used as one of the stronger comparison systems, at around one third of the inference cost.
There are still some cases where Echo makes the wrong allocation or combination decision. I’m currently spending a lot of time understanding those failures, as well as testing whether the same approach holds up on coding and agentic tasks where measuring the quality of each decision becomes much harder.
I built a chat interface (echo.tracerml.ai) and an OpenAI-compatible API (https://echo.tracerml.ai/docs/api) so the system can be tested outside the evaluation setup.
Here is a short/high level video on how it works: https://www.youtube.com/watch?v=lJFJSvOdXhg
I wrote up the evaluation methodology, individual model results, costs and current limitations here: https://echo.tracerml.ai/eval
I would love for you to try it! Especially if you hit any weird failure cases or places where the allocation looks unintuitive.", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49029215", "k": "Negative Resistance", "domain": "technology", "knowledge": "Negative Resistance", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49031409", "k": "Claude Cookbook", "domain": "technology", "knowledge": "Claude Cookbook", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49007291", "k": "PyPI Blog: Releases now reject new files after 14 days", "domain": "technology", "knowledge": "PyPI Blog: Releases now reject new files after 14 days", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49047526", "k": "The AI Productivity Illusion", "domain": "technology", "knowledge": "The AI Productivity Illusion", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49010993", "k": "Building a Tiny 3D Renderer for a Tiny Handheld", "domain": "technology", "knowledge": "Building a Tiny 3D Renderer for a Tiny Handheld", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49055752", "k": "Claude Code Cut Their System Prompt by 80%. Does That Work for Small Models Too?", "domain": "technology", "knowledge": "Claude Code Cut Their System Prompt by 80%. Does That Work for Small Models Too?", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49054594", "k": "95CSS: A Windows 95-style CSS framework", "domain": "technology", "knowledge": "95CSS: A Windows 95-style CSS framework", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49034217", "k": "The case for MUDs in modern times (2018)", "domain": "technology", "knowledge": "The case for MUDs in modern times (2018)", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49046917", "k": "What if the RAM/GPU shortage is deliberate?", "domain": "technology", "knowledge": "What if the RAM/GPU shortage is deliberate?", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49048212", "k": "Engineering management after the cost of code collapsed", "domain": "technology", "knowledge": "Engineering management after the cost of code collapsed", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49057902", "k": "NASA Fire Information Interactive Map", "domain": "technology", "knowledge": "NASA Fire Information Interactive Map", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49034292", "k": "My security camera shipped a GitHub admin token in its login page", "domain": "technology", "knowledge": "My security camera shipped a GitHub admin token in its login page", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49008001", "k": "Retry Storm Lab", "domain": "technology", "knowledge": "Retry Storm Lab", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49002235", "k": "Librrd Playground", "domain": "technology", "knowledge": "Librrd Playground", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_48995409", "k": "Firefox Containers Preview", "domain": "technology", "knowledge": "Firefox Containers Preview", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49023016", "k": "Startup founders urge U.S. government not to shut off Chinese open weight AI", "domain": "technology", "knowledge": "https://littletech.org/
https://static.politico.com/4a/bf/9c4021d8404386b0a311dcccf0...", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49036433", "k": "Government orders GitHub to remove Bluetooth-based chat app Bitchat: Jack Dorsey", "domain": "technology", "knowledge": "Government orders GitHub to remove Bluetooth-based chat app Bitchat: Jack Dorsey", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49055475", "k": "Show HN: Inflect TTS v2+ONNX, 9M/4M text-to-speech models running in the browser", "domain": "technology", "knowledge": "Show HN: Inflect TTS v2+ONNX, 9M/4M text-to-speech models running in the browser", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49040741", "k": "Opus 5 is currently #1 on Artificial Analysis Intelligence Leaderboard", "domain": "technology", "knowledge": "Opus 5 is currently #1 on Artificial Analysis Intelligence Leaderboard", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49057716", "k": "Introduction to Microsoft Excel 1990 with Jan Brehm", "domain": "technology", "knowledge": "Introduction to Microsoft Excel 1990 with Jan Brehm", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49055436", "k": "Sequentialising Parallel Moves", "domain": "technology", "knowledge": "Sequentialising Parallel Moves", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49040296", "k": "Postgres LISTEN/NOTIFY actually scales", "domain": "technology", "knowledge": "Postgres LISTEN/NOTIFY actually scales", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49015898", "k": "Agatha Christie's Disappearance: Amnesia, Suicide, or Despair?", "domain": "technology", "knowledge": "Agatha Christie's Disappearance: Amnesia, Suicide, or Despair?", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49049814", "k": "California Historical Radio Society", "domain": "technology", "knowledge": "California Historical Radio Society", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49045040", "k": "ARC-AGI Leaderboard", "domain": "technology", "knowledge": "ARC-AGI Leaderboard", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49055400", "k": "Show HN: Pg_stat_ch, a Postgres extension to export every metric to ClickHouse", "domain": "technology", "knowledge": "Show HN: Pg_stat_ch, a Postgres extension to export every metric to ClickHouse", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49052025", "k": "Digital Intelligence: Why Humanity Created What It Does Not Understand", "domain": "technology", "knowledge": "Digital Intelligence: Why Humanity Created What It Does Not Understand", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49053986", "k": "Show HN: What 180k words look like as a temporal knowledge graph (Oz series)", "domain": "technology", "knowledge": "The graph is free to explore and requires no registration.
SynapTale builds a model of a story as a temporal graph made up of nodes (entities) and edges (their actions and relationships).\nThe graph is not a visualization of the wiki. The wiki, timelines, relationship histories, and analytics are projections of the graph.
The current demo contains 232 entities, 1,852 edges, and a snapshot of the story’s state at every chapter. By chapter 100, it still remembers a promise made in chapter 8 and turns the story into a set of source-verifiable facts.
The most interesting things can be found in the graph itself and in the Analytics tab. A few things I found:
1. The character with the highest kill count is the Tin Woodman—the same character who cries over a beetle he accidentally crushed. Dorothy comes second, with three killing events.\n2. Dorothy never deceives anyone during the first 100 chapters of the series.\n3. The Scarecrow’s debt to the stork has remained active for 92 chapters, starting in chapter 8.\n4. The Cowardly Lion ranks third by number of threats.\n5. The first 100 chapters contain 60 secrets and 254 dialogue events.
Technical details
1. Five different multi-agent pipelines combining LLMs and NLP: a prescan, ontology construction, chapter-by-chapter graph extraction, retrospective validation over spans of dozens of chapters, and a linguistic prescan for speech profiles and linguistic edges.
2. A living story needs a living graph. It has to account for time, because entities and the relationships between them evolve. A simple is_active field is not enough.
I ended up with three types of edges:
event: an instantaneous action;\nidentity: a fact;\nstate: a persistent action whose termination requires justification and a supporting quote from the text.
The vast majority of edges are events and end in the same chapter in which they began. This allows the system to scale well, since only a minority of state and identity edges remain continuously active.
3. Ontology. You cannot simply ask an LLM to extract entities and relationships into a graph. With every chapter, even the smartest model will keep inventing unimportant fields, creating new aliases for existing fields, and representing the same fields inconsistently.
Before extracting the graph, the system therefore performs an ontology scan across the entire story. It captures story-specific entity and edge types, along with their fields and descriptions.
4. Epistemics. Events are only one part of a story. It is also important to understand how information is distributed, which is difficult to represent using event edges alone.
I addressed this by introducing a new node type: epistemic nodes, which capture different entities’ perspectives on the same fact. Subtle hints can still be missed, the system is not yet perfect in this area.", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49055305", "k": "Watching Go's new garbage collector move through the heap", "domain": "technology", "knowledge": "Watching Go's new garbage collector move through the heap", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49038045", "k": "Unitree As2-W", "domain": "technology", "knowledge": "Unitree As2-W", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49053805", "k": "Show HN: ZWPlayer – Free player with interactive annotations and subtitle search", "domain": "technology", "knowledge": "Show HN: ZWPlayer – Free player with interactive annotations and subtitle search", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49042528", "k": "SpaceX Starship Flight 13 livestream [video]", "domain": "technology", "knowledge": "SpaceX Starship Flight 13 livestream [video]", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49056968", "k": "Registration of new accounts to the Arch AUR reopened with hardened requirements", "domain": "technology", "knowledge": "Registration of new accounts to the Arch AUR reopened with hardened requirements", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49004464", "k": "Marimo now runs in PyCharm", "domain": "technology", "knowledge": "Marimo now runs in PyCharm", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49025766", "k": "Why Sony can't bring back its classic Walkman models", "domain": "technology", "knowledge": "Why Sony can't bring back its classic Walkman models", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49050152", "k": "Tile's security is so bad it's a feature for stalkers", "domain": "technology", "knowledge": "Tile's security is so bad it's a feature for stalkers", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49012604", "k": "Book Corners: Community map of neighborhood book exchange spots", "domain": "technology", "knowledge": "Book Corners: Community map of neighborhood book exchange spots", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49047707", "k": "Scanwheel is a drum style mechanical television you can build yourself", "domain": "technology", "knowledge": "Scanwheel is a drum style mechanical television you can build yourself", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49054268", "k": "Agentic test processes, LLM benchmarks, and other notes on agentic coding", "domain": "technology", "knowledge": "Agentic test processes, LLM benchmarks, and other notes on agentic coding", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49038060", "k": "Be skeptical of OpenAI's rogue hacker agent story", "domain": "technology", "knowledge": "Be skeptical of OpenAI's rogue hacker agent story", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49034747", "k": "Em dashes are amazing", "domain": "technology", "knowledge": "Em dashes are amazing", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49032095", "k": "The PImpl idiom and the C++26 std:indirect type", "domain": "technology", "knowledge": "The PImpl idiom and the C++26 std:indirect type", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49012337", "k": "Teardown: The BMW / Harman IDC23H Infotainment Unit (B423)", "domain": "technology", "knowledge": "Teardown: The BMW / Harman IDC23H Infotainment Unit (B423)", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49050218", "k": "ESP32-C6 Power Consumption: Arduino vs. Zephyr vs. ESP-IDF Comparison", "domain": "technology", "knowledge": "ESP32-C6 Power Consumption: Arduino vs. Zephyr vs. ESP-IDF Comparison", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49054958", "k": "Kaappi – r7rs scheme written in Zig", "domain": "technology", "knowledge": "Kaappi – r7rs scheme written in Zig", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49052552", "k": "Show HN: Hubo – two agents that implement and review code until they agree", "domain": "technology", "knowledge": "Show HN: Hubo – two agents that implement and review code until they agree", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49056808", "k": "An OpenAI model left notes about how to evade containment; we need more details", "domain": "technology", "knowledge": "An OpenAI model left notes about how to evade containment; we need more details", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49024938", "k": "Kimi K3 exploited the latest Redis server", "domain": "technology", "knowledge": "https://xcancel.com/fried_rice/status/2080059356322918777", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49057318", "k": "Google slapped with $1B fine under landmark EU digital law", "domain": "technology", "knowledge": "Google slapped with $1B fine under landmark EU digital law", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49057254", "k": "Bond market anxiety is growing over AI capex budgets", "domain": "technology", "knowledge": "Bond market anxiety is growing over AI capex budgets", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49056732", "k": "Resonance: Microphysics Meets the Cosmological Constant", "domain": "technology", "knowledge": "Resonance: Microphysics Meets the Cosmological Constant", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49052181", "k": "Show HN: Rudoc – a 4.5MB Rust document converter", "domain": "technology", "knowledge": "I built Rudoc because I wanted a smaller tool for simple file convertion workflow. (Pandoc was so heavy for me) I once wanted to try Pandoc, but after I downloaded it from GitHub, I found that it is around 70 MB. It also needed a lot of dependencies.
Rudoc is not a substitue of Pandoc, it supports limited set of formats (only txt, md, html, typ, docx, pdf, pptx, xml, json) and doesn't support advanced formatting. The most common workflow is just turn a markdown file to a html. PDF generation can work without Typst, but PDF input conversion requires Typst as a backend. It has no other runtime dependencies and can run as a single binary, by the way.
It is pretty easy to use, just use the format of\nrudoc input.file output.file\nOne sample usage is like this:\n```powershell\nrudoc OhPDF_ROADMAP.md.txt r.md\n[rudoc] OhPDF_ROADMAP.md.txt → r.md (txt→md)\n[rudoc] done in 52ms (6286 bytes)\nrudoc r.md r.html\n[rudoc] r.md → r.html (md→html)\n[rudoc] done in 4ms (7363 bytes)\n```
Some complex formats (such as pptx) still have limited compatibility. I planned to add standard error messages and better format compatibilities. I'd be pleased to see feedback or issues on my project.
QA:\nWhy not Pandoc?\nPandoc is a great tool, but sometimes we only need a convenient little one to handle common conversion tasks.
How compatible is it to docx/pptx?\nCurrently it supports basic text conversion, and I'm trying to make it be able to keep advanced formatting.", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49054819", "k": "Registration Open for OS/2 and ArcaOS Warpstock 2026 in San Antonio, Texas, USA", "domain": "technology", "knowledge": "Registration Open for OS/2 and ArcaOS Warpstock 2026 in San Antonio, Texas, USA", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49054801", "k": "Betlang – A programming language detection model in 50kb", "domain": "technology", "knowledge": "Betlang – A programming language detection model in 50kb", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49034673", "k": "Programming language file extensions that match ISO 3166-1 alpha-2 country codes", "domain": "technology", "knowledge": "Programming language file extensions that match ISO 3166-1 alpha-2 country codes", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49022137", "k": "Fields Medals 2026", "domain": "technology", "knowledge": "Fields Medals 2026", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49052937", "k": "Dht-spy – BitTorrent DHT telemetry", "domain": "technology", "knowledge": "Dht-spy – BitTorrent DHT telemetry", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49036765", "k": "Codeberg Divides", "domain": "technology", "knowledge": "Codeberg Divides", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49033240", "k": "IRGC claims it destroyed Amazon's Bahrain data center", "domain": "technology", "knowledge": "IRGC claims it destroyed Amazon's Bahrain data center", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49049066", "k": "Rauno's Field Notes #2", "domain": "technology", "knowledge": "Rauno's Field Notes #2", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49054696", "k": "Codegraph", "domain": "technology", "knowledge": "Codegraph", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49052798", "k": "Htmx 4: The Game", "domain": "technology", "knowledge": "Htmx 4: The Game", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49031796", "k": "Flux 3", "domain": "technology", "knowledge": "Flux 3", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49022911", "k": "Show HN: Palmier Pro – Open-source macOS video editor built for AI", "domain": "technology", "knowledge": "Hi HN, we are Marcos and Harrison, cofounders of Palmier (https://palmier.io). We are building Palmier Pro, an open source macOS video editor, with built-in AI generation and a local MCP server that connects to your agent. Here are a few demos:
- Making some AI transitions: https://www.youtube.com/watch?v=hbM_-eR1GX4
- Multicam editing with Codex: https://www.youtube.com/watch?v=SjS2q2LT1q8
- Cutting long form clips into shorts: https://www.youtube.com/watch?v=PR66eN2ouuQ
We built Palmier Pro as an internal tool when we were making AI launch videos for other startups. The main problem it solved in the beginning was the back-and-forth between AI generation platform and video editor. The iteration loop was awkward: AI videos → download → import to editor → edit → realize we need to change the AI video → repeat. So we built a minimal video editor where we could let Claude generate AI videos inside the editor.
As we gave more and more tools to the agent, we wanted to push to see what else agents can do in the video editing space. So today, your Claude/Codex can:
- Manage projects inside Palmier Pro
- Import media from a public URL or filesystem to the project, and organize them in folders
- Search media (by embedding footages using SigLIP2 running locally)
- Edit the timeline (tracks/clips/keyframes operations)
- Generate images, videos, sound effects, captions, music
- Export videos
There are two ways for LLMs to interact with the editor: by connecting to the local MCP server, or using the in-app chat. Both use the same tools and APIs exposed by the video editor.
We have seen people using MCP server to connect to their own workflow to automate massive-scale video editing (e.g. given this same podcast style, replicate it with other footages that I have). We have also seen people using the in-app chat where it lives closer to the editor UI, with lower latency for faster iteration.
We don't believe that AI is going to replace human creativity (nor should it), but where it can really help is in automating the gruntwork in video editing. Most work in the editing workflow is more mechanical than creative. Our vision is to build a video editor where AI can enable more individuals to create.
Throughout our experiments, AI is not very good at creative editing, but given a pattern (transcription-based, beat-based), it can do a decent job at rough cut.
We built Palmier Pro in Swift as an MVP because we wanted better performance and minimal dependencies (no nodejs/webview), with some native macOS APIs like SpeechAnalyzer and CoreML for us to run some models locally. We use SpeechAnalyzer for local transcriber, SigLip2 to embed video frames, beat_this for beat detection, Silero VAD for silence detection, all running locally. The tradeoff is that we don't support Linux or Windows at the moment.
Palmier Pro is open source and free to try out (macOS 26 only at the moment, though - we'll support more platforms over time, but for now are focusing on iterating the core product).
No login required except for the AI generation features, which route requests to our backend. We are offering free credits on sign up so you can try out the AI generation as well.
We'd love to hear your feedback!", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49054602", "k": "Interlock: Proving a secret moved, instead of guessing that it might", "domain": "technology", "knowledge": "Interlock: Proving a secret moved, instead of guessing that it might", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49056424", "k": "No Stack Overflow, No Autocomplete: What Coding Felt Like in the 80s", "domain": "technology", "knowledge": "No Stack Overflow, No Autocomplete: What Coding Felt Like in the 80s", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49033110", "k": "Future euro banknote design proposals", "domain": "technology", "knowledge": "https://www.ecb.europa.eu/euro/banknotes/future_banknotes/ht...", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49053255", "k": "GNU Hurd gets 9pfs, OpenNTPD, dynamic /dev/ entries, and more", "domain": "technology", "knowledge": "GNU Hurd gets 9pfs, OpenNTPD, dynamic /dev/ entries, and more", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49034868", "k": "Half-Life 2 running natively on HaikuOS", "domain": "technology", "knowledge": "Half-Life 2 running natively on HaikuOS", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49054561", "k": "Kimi K3 after a week of real project use: what works, what doesn't", "domain": "technology", "knowledge": "Kimi K3 after a week of real project use: what works, what doesn't", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49057046", "k": "Canada remains anti-nuke, but its potential path to the bomb is getting shorter", "domain": "technology", "knowledge": "Canada remains anti-nuke, but its potential path to the bomb is getting shorter", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49052508", "k": "Littleman Programming", "domain": "technology", "knowledge": "Littleman Programming", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49054525", "k": "Juice – non-essential visual, audio, and haptic effects to enhance UX", "domain": "technology", "knowledge": "Juice – non-essential visual, audio, and haptic effects to enhance UX", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49053520", "k": "I scanned my AI agent framework for destructive/consequential actions, and wow", "domain": "technology", "knowledge": "I scanned my AI agent framework for destructive/consequential actions, and wow", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49000864", "k": "LG to ban residential proxies from smart TV apps", "domain": "technology", "knowledge": "LG to ban residential proxies from smart TV apps", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49054197", "k": "India's 'Cockroach' Protesters Declare Rare Victory over Government", "domain": "technology", "knowledge": "India's 'Cockroach' Protesters Declare Rare Victory over Government", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49039395", "k": "Gsxui – Shadcn-style components for Go", "domain": "technology", "knowledge": "Gsxui – Shadcn-style components for Go", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49052302", "k": "Western Europe, State Formation, and Genetic Pacification", "domain": "technology", "knowledge": "Western Europe, State Formation, and Genetic Pacification", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49020545", "k": "Show HN: I simulated closing the Strait of Hormuz on real oil trade data", "domain": "technology", "knowledge": "OP here: I created this visualization tool as the byproduct of a supply chain class I taught at Columbia. The pedagogical exercise grew into a full blown visualization and paper about global oil trade.
The model:\nThe mechanics are the same as the financial network Eisenberg-Noe: Instead of banks, every country consumes oil interconnected via bilateral trading. Shocks propagate throughout the network, depleting oil reserves when bottleneck nodes (such as the Strait of Hormuz) are blocked.
Insights:\nThe interesting part is the mechanics of how the crisis unfolds: for example, France receives 0 oil from Hormuz directly, yet their reserves are depleted faster because other countries reactively increase their safety oil stock, increasing oil price, making stockouts more expensive for everyone.
The model also gives price dynamics which are interesting on their own: the price increase is not immediate, it follows sequentially as countries reserves deplete.
Some caveats:\n1. For producer nodes, depletion means their export slack is reduced/exhausted.\n2. No sanctioned trade (UN Comtrade data)
Technical Details:\nThe visualization is 600 lines of flask plus js frontend (LLM assisted visualization with ground-truth matching the original numerical exercise of the paper)
Paper with proofs/theory:\nhttps://arxiv.org/abs/2607.17491", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49055279", "k": "GrapheneOS duress PIN could land a man in prison", "domain": "technology", "knowledge": "GrapheneOS duress PIN could land a man in prison", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49048961", "k": "Re: I Stopped \"Creating Content\"", "domain": "technology", "knowledge": "Re: I Stopped \"Creating Content\"", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49052081", "k": "AI isn't killing consulting. It's killing time as a proxy for value", "domain": "technology", "knowledge": "AI isn't killing consulting. It's killing time as a proxy for value", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49038481", "k": "DuckPGQ – A DuckDB community extension for graph workloads", "domain": "technology", "knowledge": "DuckPGQ – A DuckDB community extension for graph workloads", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49037031", "k": "The front end framework for correctness: built on Effect, architected like Elm", "domain": "technology", "knowledge": "The front end framework for correctness: built on Effect, architected like Elm", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49056892", "k": "The perils of parsing type inference declarations in C", "domain": "technology", "knowledge": "The perils of parsing type inference declarations in C", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_48997828", "k": "Some more things about Django I've been enjoying", "domain": "technology", "knowledge": "Some more things about Django I've been enjoying", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49049724", "k": "Some combinatorial applications of spacefilling curves", "domain": "technology", "knowledge": "Some combinatorial applications of spacefilling curves", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49053135", "k": "Triton: DirectX 11 driver for QEMU", "domain": "technology", "knowledge": "Triton: DirectX 11 driver for QEMU", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49056855", "k": "AI Chatbots Know How to Make Deadly Biological Weapons. Some Will Teach You", "domain": "technology", "knowledge": "AI Chatbots Know How to Make Deadly Biological Weapons. Some Will Teach You", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49056294", "k": "Emacs Writing Machine", "domain": "technology", "knowledge": "Emacs Writing Machine", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49056286", "k": "Robots programmed by kids to speak their Indigenous languages", "domain": "technology", "knowledge": "Robots programmed by kids to speak their Indigenous languages", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49054638", "k": "MIT to become hotbed of AI video surveillance", "domain": "technology", "knowledge": "MIT to become hotbed of AI video surveillance", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49034342", "k": "I Tried Building a Real App with AI. It Took a Year", "domain": "technology", "knowledge": "I Tried Building a Real App with AI. It Took a Year", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_48994440", "k": "Latest Airbus single aisle aircraft innovations", "domain": "technology", "knowledge": "Latest Airbus single aisle aircraft innovations", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49056753", "k": "Wildfire Forces Evacuation of NASA's Deep Space Network Complex in Spain", "domain": "technology", "knowledge": "Wildfire Forces Evacuation of NASA's Deep Space Network Complex in Spain", "authority": 3.5, "_expert": "expert_tech"} {"id": "hn_49051282", "k": "Apple seeks U.S. approval to buy memory chips from blacklisted CXMT", "domain": "technology", "knowledge": "Apple seeks U.S. approval to buy memory chips from blacklisted CXMT", "authority": 3.5, "_expert": "expert_tech"} {"id": "arxiv_9904d42c25", "k": "3D-Aware VLMs with Implicit and Explicit Geometries", "domain": "science", "knowledge": "Despite rapid progress, most existing vision-language models (VLMs) built from 2D visual inputs often struggle when handling various 3D tasks that require fine-grained spatial understanding and reasoning. To bridge this gap, we present VLM-IE3D, a unified framework that enhances the 3D spatial awareness of VLMs by equipping them with both implicit and explicit 3D geometries learned from RGB videos. Our VLM-IE3D introduces Implicit Geometry Tokens (IGTs) that capture high-level geometric priors from input videos, as well as complementary Explicit Geometry Tokens (EGTs) that encode detailed geometric structures from reconstructed 3D attributes. On top of that, VLM-IE3D comes with a 3D-aware adapter that effectively fuses the two types of geometric representations with 2D visual cues. This RGB-only design injects strong 3D inductive biases for fine-grained spatial understanding and reasoning without requiring any additional 3D inputs. Extensive experiments show that VLM-IE3D achieves superior performance consistently across various 3D tasks including 3D video detection, 3D visual grounding, 3D dense captioning, and spatial reasoning. Code and models are available at https://github.com/Vegetebird/VLM-IE3D.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_d0dcfcba98", "k": "GraphVid: Interactive Graph-Controllable Video Generation", "domain": "science", "knowledge": "Controllable video generation remains challenging due to the difficulty of specifying precise multi-object interactions using text prompts or motion-control inputs that primarily constrain pixel movement. In practice, trajectory-based control often requires users to draw accurate tracks for multiple objects, which scales poorly with scene complexity and becomes ambiguous under occlusion or overlap. To enable flexible yet precise multi-subject control, we introduce $\\textbf{GraphVid}$, a graph-conditioned image-to-video generation model that enables interactive control through structured interaction graphs. We further curate $\\textbf{GraphVid-Bench}$, a large-scale interaction-centric video dataset with structured relational annotations to enable training of interaction-aware video generation models. Despite using substantially less training data and fewer trainable parameters than prior motion-control methods, GraphVid delivers strong controllability and video quality. Compared with Motion-I2V, GraphVid reduces FID by up to 39.9% and FVD by 37.6%, while improving PSNR (9.87=>15.98) and SSIM (0.38=>0.61). Our results highlight the potential of structured semantic interfaces as a powerful paradigm for controllable video generation.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_a05a246486", "k": "Barzilai-Borwein Fails Superlinear Convergence on an Open Set of Quadratics for Every Dimension $n\\geq 4$", "domain": "science", "knowledge": "Barzilai--Borwein (BB) method has shown strong practical performance in continuous optimization, yet its convergence dynamics remains poorly understood. In particular, a central unresolved question is whether BB converges superlinearly for almost every strictly convex quadratic problem and initialization. We provide a negative answer to this question. Specifically, for every finite dimension $n\\geq4$, we construct a nonempty open, hence positive-Lebesgue-measure, family of strictly convex quadratic problems and initial points for which the long Barzilai--Borwein method (BB1) converges but cannot converge root-superlinearly. More precisely, with the explicit constants $ρ_{\\min}=10^{-6},ρ_{\\max}=0.61$, every spectral component of the gradient is bounded above and below by the corresponding geometric sequence. Consequently, the gradient norm and the energy norm of the error satisfy two-sided geometric estimates with the same rates, while the objective gap satisfies the corresponding estimates with squared rates. In particular, all three quantities are bounded below by geometric sequences, ruling out superlinear convergence. The construction is highly nontrivial, based on a computer-assisted proof of a nonresonant, attracting seven-cycle of the projectivized BB dynamics in dimension four.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_b82ad00c90", "k": "Synthetic data generation framework for quality control automation in gravure printing", "domain": "science", "knowledge": "Quality control in printing, particularly in rotogravure printing, still depends on slow, costly, and subjective manual inspection. Automated surface defect detection is critical for maintaining high-quality standards in rotogravure printing. Deep learning models give prospects for automation. However, training robust deep learning models, such as YOLO or Vision Transformers, is heavily hindered by the extreme scarcity of real-world industrial defects images. To overcome this limitation, this paper introduces a novel synthetic data generation framework tailored for rotogravure printing quality control. The proposed pipeline automatically generates high-fidelity images of specific printing defects (creases, streaks, misregistration, etc.) and outputs corresponding bounding boxes and annotations. To validate the framework, a synthetic dataset of 7533 images was generated and used to train the state-of-the-art object-detection model RFDETR. Experimental results demonstrate that the model trained on our synthetic data achieves a Mean Average Precision (mAP) of 80.9\\% on real industrial testing samples. This framework provides a zero-cost, rapid-deployment solution for automating defect inspection in printing lines without requiring massive manual data collection.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_928c2671ab", "k": "Beyond Sufficiency: Time Series Explanation with Counterfactual Necessity", "domain": "science", "knowledge": "Faithful explanations of time-series classifiers should identify subsequences that are not only sufficient to preserve a black-box model's prediction, but also necessary for maintaining it. However, existing sufficiency-oriented methods can assign high importance to spurious subsequences that support the prediction without being essential to the model's decision. We introduce \\textbf{TimePNS}, a necessity-aware framework for time-series explanation. Inspired by Pearl's counterfactual notion of necessity, TimePNS assesses whether a temporal factor is necessary by intervening on it and measuring whether the original prediction is disrupted. The framework adopts a two-stage design. Stage I learns an identifiable causal generative process together with a sufficiency-oriented explanation mask. Stage II performs counterfactual interventions on temporal factors to derive necessity signals, which supervise a temporal gate that refines the initial explanation by suppressing non-essential components and emphasizing counterfactually necessary ones. Experiments on synthetic and real-world time-series benchmarks show that TimePNS more accurately identifies decision-critical subsequences and consistently improves sufficiency-necessity trade-offs over strong baselines.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_821521b72a", "k": "Unsupervised Consensus-Based Anomaly Detection for Spatiotemporal Malaria Incidence in Ghana", "domain": "science", "knowledge": "A consensus anomaly detection framework was applied to monthly malaria surveillance data from Ghana (2014-2023) to identify atypical transmission patterns. Anomalies were highly structured in space and time. Ashanti and Northern Regions accounted for most recurrent anomalies, with persistent hotspots at Tamale, Kumasi, and Accra. A key finding was the spatial distinction between anomaly burden (cumulative cases during anomalous periods) and anomaly frequency (persistence of unusual behaviour). Tamale had the highest burden during anomalies, whereas the highest anomaly rates clustered in Ashanti districts, showing that high-burden areas are not necessarily those with the most frequent anomalous transmission. Anomalous months formed a statistically distinct group, with much higher case counts (Cohen's $d = 3.252$) and large seasonal deviations ($d > 1.2$) compared with normal months. Malaria burden alone provides an incomplete picture of transmission dynamics. By distinguishing where malaria is most prevalent from where transmission behaves most unusually, this framework can strengthen surveillance, prioritise investigations, and support targeted control strategies.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_09a5d1b41d", "k": "Beyond Sycophancy: Structured Resistance and Compliance in LLM Moral Reasoning", "domain": "science", "knowledge": "Building socially calibrated large language models, which can learn from others without simply yielding to them, requires more than reducing sycophancy as a one-dimensional failure mode. Models must distinguish when to incorporate others' perspectives from when to maintain a well-grounded moral judgment. We study the broader resistance-compliance process governing this distinction. Across three studies, we show that models' judgment revision is structured along three dimensions that parallel classic phenomena in human social psychology: the distance between an incoming view and the model's initial position, the source attribution of that view, and the coalition structure supporting it. Models are generally more receptive to nearby positions, more influenced by views presented as their own prior judgments, and differently responsive to group pressure. These findings recast sycophancy as one expression of a broader judgment-updating process shaped by social influence. Our framework provides a principled basis for distinguishing constructive belief revision from sycophantic compliance, thereby supporting better alignment in morally consequential interactions.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_3451f3466f", "k": "OpenForgeRL: Train Harness-native Agents in Any Environment", "domain": "science", "knowledge": "Modern AI agents rely on elaborate inference harnesses such as Claude Code, Codex, and OpenClaw to drive multi-turn reasoning, tool use, and access to external systems. While powerful, these complex harnesses also make agents hard to train end-to-end with open infrastructure, whose SFT/RL stacks cannot natively express stateful, multi-process harness inference. To address this, we present OpenForgeRL, an open-source framework for training harness-based agents end-to-end in diverse environments. OpenForgeRL achieves this with a lightweight proxy that serves the harness's model calls while recording them as training data for a standard RL codebase (e.g., veRL), and a Kubernetes orchestrator that runs each rollout in its own remote container, together enabling training on any harness in any environment at scale. By decoupling training and inference, OpenForgeRL allows researchers to easily train, study, and improve agents directly in the real harnesses and environments they are deployed with. We validate our framework across diverse, complex harnesses and environments, spanning tool/claw-based agents and multimodal GUI browser- and computer-use agents. Using only hundreds to a few thousand tasks, OpenForgeClaw reaches 31.7 pass^3 and 55.9 pass@3 on ClawEval and 33.7 on QwenClawBench. OpenForgeGUI reaches 37.7 on OSWorld-Verified, 63.0 on Online-Mind2Web, and 72.3 on WebVoyager. Both outperform open baselines of similar size on nearly all benchmarks, and in the GUI setting match or surpass models several times larger. Beyond benchmarks, we analyze how harness choice (e.g., ZeroClaw, OpenClaw, Codex) and RL shape agent behavior. We find that some harnesses are substantially harder to learn than others, and that RL improves agentic reliability, such as self-verification, tool coverage, and completing multi-step plans, though critical abilities such as error recovery remain weak.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_bc6eda1cf7", "k": "Visual Contrastive Self-Distillation", "domain": "science", "knowledge": "On-policy self-distillation (OPSD) is promising as it removes the external teacher required by on-policy distillation (OPD), yet it still needs asymmetric information between teacher and student to ensure that the self-teacher provides a stronger learning signal than the student. Existing methods create this asymmetry either through privileged answers or visual evidence. We ask whether both can be removed, yielding a simpler form of OPSD driven purely by input conditioning. For this purpose, we propose Visual Contrastive Self-Distillation, namely VCSD, which converts image-content removal into an on-policy self-distillation signal. At each student-generated response prefix, the EMA teacher produces two next-token distributions under the same prompt and prefix -- one conditioned on the original image and the other on a content-erased control. Their token-wise log-probability difference highlights candidates whose likelihood is specifically increased by the instance-level visual content. We use this contrast to sharpen the teacher's original-image distribution within its plausible support, and distill the resulting full-distribution target into the student. Using ViRL39K dataset, VCSD consistently outperforms matched OPSD across Qwen3-VL and Qwen3.5 models. For example, on Qwen3-VL, it improves the seven-benchmark aggregate from $62.27\\% \\rightarrow 67.04\\%$ at 2B, $71.30\\% \\rightarrow 73.16\\%$ at 4B, and $72.51\\% \\rightarrow 76.26\\%$ at 8B. Furthermore, VCSD requires no external teacher, privileged answers, visual evidence signals, reasoning traces, or additional inference-time cost.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_e8e8149dc2", "k": "MIRROR: Learning from the Other View for Multi-Modal Reasoning", "domain": "science", "knowledge": "Unlike large language models (LLMs) that exhibit strong reasoning capabilities, vision-language models (VLMs) struggle with visual reasoning, even on geometry problems that admit equivalent text, diagram, and combined diagram+text views. We show that these views often elicit different behaviors: a model may solve a problem from text but fail on the corresponding diagram, or succeed visually while failing textually. This inconsistency suggests that different views expose complementary reasoning paths and failure modes that standard multimodal post-training does not fully exploit. To study and exploit this phenomenon, we construct ODA-Data, a high-quality paired multimodal geometry dataset with text-dominant, image-dominant, and combined image+text views of the same problems, together with splits for training and evaluating modality-dependent reasoning behaviors. We then develop Modality-Informed Reciprocal Reasoning Optimization (MIRROR), a reinforcement learning approach for improving multimodal reasoning via self supervision. For each problem, MIRROR evaluates the model under all views, selects the best-performing view as a teacher, and trains other views with a reverse-KL objective towards the teacher. Across reasoning benchmarks that evaluate on geometry problems, MIRROR improves over standard RL and yields more accurate and consistent behavior across modalities", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_e2b0eed731", "k": "The Boundaries of Automation: A Theory of Persistent Human Participation", "domain": "science", "knowledge": "The rapid progress of AI has intensified the long-standing pursuit of automation: replacing human participation with algorithms wherever possible. Implicit in this pursuit is the assumption that humans remain in the loop only because current AI systems are not yet sufficiently capable. This paper challenges that assumption. Rather than asking how far automation can extend, we ask where its conceptual limits lie and argue that human participation may persist even with highly capable AI systems for three distinct reasons. Technical or complementarity grounds arise when humans contribute capabilities or perspectives unavailable to AI. Normative or developmental grounds arise when participation itself is valuable for human agency or learning. Most importantly, emergence grounds arise from target emergence: in some activities, the target is not fully specified in advance but instead emerges through the interaction itself. In these cases, human participation is not merely a means of improving execution but is constitutive of the target being produced. Human--AI co-construction, understood as the joint production of outcomes by humans and AI systems, is therefore not simply a temporary response to imperfect AI, but a persistent feature of activities whose objectives emerge through participation. This perspective has important implications for the limits of automation and for the design, evaluation, and ethics of future AI systems.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_f28a755734", "k": "From Resource Flow to Executable Tests: Petri-Net-Guided LLM Test Generation for Concurrent Stateful Rust APIs", "domain": "science", "knowledge": "Concurrent stateful library APIs expose behavior through evolving resource ownership, lifecycle states, and competing interleavings. Large language models can synthesize executable Rust tests, but their outputs often violate API preconditions, remain shallow, or reduce concurrency to accidental sequential traces. Conversely, model-based and systematic testing techniques provide semantic control but commonly require substantial handwritten code to turn abstract scenarios into executable tests. This paper addresses the gap between formal scenario design and low-cost test concretization. We present a Petri-net-guided methodology for test generation over concurrent stateful Rust APIs. The method represents API resources, lifecycle conditions, and causal dependencies as colored tokens and transitions; derives legal deep-state, near-legal, and partial-order concurrent scenarios; and uses these scenarios as a constrained intermediate representation for LLM-based code synthesis. A local-faithfulness contract and structural repair loop preserve the modeled intent during concretization, while Petri-guided schedule shaping prioritizes high-conflict concurrency skeletons for systematic exploration. A layered semantic oracle then distinguishes synthesis failures from violations of the target API's expected behavior.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_4c485b177c", "k": "ElasticTTT: Prior-Preserving Test-Time Tuning for Video Editing", "domain": "science", "knowledge": "Test-Time Tuning (TTT) on pretrained diffusion models has emerged as a powerful paradigm for video editing. However, there exists a foundational mismatch between the distribution-mapping nature of generative models and the single-point optimization of standard TTT. In this paper, we demonstrate that this mismatch triggers \\textit{Prior Collapse}, a degenerate state where the model discards the text conditions and spatial latents, collapsing generations to the source video, or entangling the features of distinct regions. To resolve this, we propose \\textbf{ElasticTTT}, a novel framework that preserves the prior generative distribution and rescues generative elasticity. Specifically, we propose \\textit{Target Distribution Regularization} to prevent sharp memorization minima, \\textit{Contrastive CFG} to guide inference away from source biases, and \\textit{Asynchronous Noise Schedule} to preserve unedited regions. Extensive evaluations, supported by theoretical analysis, demonstrate that ElasticTTT successfully preserves the generative prior of the base model, achieving state-of-the-art performance on one-shot video editing.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_c5e6e4ab00", "k": "GS-Agent: Creating 4D Physical Worlds With Generative Simulation", "domain": "science", "knowledge": "Creating dynamic and physically realistic 4D worlds from natural language descriptions is both fascinating and challenging. Traditional computer graphics methods rely on manual creation, requiring extensive human effort to fine-tune materials, motions, and visual fidelity. Recent advances in generative foundation models have sparked interest in learning to generate such 4D worlds from large-scale data; however, existing methods still struggle to ensure physical plausibility and controllability. In this work, we take a different path by leveraging foundation models to construct an agentic system that emulates how humans traditionally create 4D worlds, yet automates the entire process. We present GS-Agent, an end-to-end multi-agent framework that integrates physics engines in the loop to generate realistic, dynamic, and controllable 4D physical worlds from natural language. Inspired by how humans build 4D worlds, GS-Agent decomposes the task into entity management, covering 3D asset curation, material tuning, placement, and motion control, and rendering configuration, including camera and lighting manipulation. Multiple agents with distinct expertise interact with the physics engine via code, seek multimodal feedback, and collaborate to iteratively construct 4D worlds that align with the given descriptions. Experimental results show that GS-Agent effectively converts natural language into diverse and physically plausible 4D worlds exhibiting rich interactions among liquids, deformable objects, and rigid bodies, while achieving cinematic camera and lighting control. We envision GS-Agent as a foundation for a new paradigm in 4D world generation, empowering creative content creation and physical AI. Project page at https://umass-embodied-agi.github.io/gs-agent/", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_dd58d22256", "k": "Same Dangerous Objective, Opposite Advice: Direct Exposure versus Multi-Agent Mediation", "domain": "science", "knowledge": "Even a current high-capability LLM can appear safer when shown a dangerous objective directly than when other agents transform and relay its direction. Using OpenAI's gpt-5.6-sol model alias, we test 25 pre-specified mirrored trade-off profiles. Direct exposure to an objective authorizing concealment, fabrication, and pressure produced advice net opposed to its target. After an Id and Censor transformed the same objective into affect and a constraint-rewritten, target-bearing intention, the user-facing Superego---which saw the preferred direction but not the raw objective, its manipulative clauses, or its source---produced advice net aligned with the target. This behavioral reverse shift is consistent with the model recognizing or distrusting the manipulative motive, although we do not identify its internal mechanism. The second result exposes a compositional safety gap: a current high-capability model can be used as the user-facing component of an automated, multi-stage workflow serving an explicitly manipulative objective. The workflow can keep the raw instruction, its manipulation-authorizing clauses, and its provenance outside the downstream model's context while preserving the objective's target direction. A user with endpoint-only access likewise cannot directly inspect those upstream messages including the objective.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_b86f20bd17", "k": "Improved lower bounds for the Shannon capacity of odd cycles", "domain": "science", "knowledge": "The Shannon capacity $Θ(G)$ of a graph $G$ quantifies the maximum rate at which information can be transmitted with zero error over a noisy channel. It is lower bounded by $α(G^d)^{1/d}$ for any $d$, where $α(G^d)$ is the independence number of the $d$-th strong power of $G$. We construct independent sets of size $134753$ in $C_7^{10}$, $21909$ in $C_{11}^{6}$, and $62530$ in $C_{13}^{6}$, improving the best known lower bounds for the Shannon capacity of these graphs to $Θ(C_7)\\geq 134753^{1/10}>3.258020$, $Θ(C_{11})\\geq 21909^{1/6}>5.289773$, and $Θ(C_{13})\\geq 62530^{1/6}>6.300109$. We also improve the best known lower bounds on the independence numbers of several individual strong powers of odd cycles that do not improve the Shannon capacity lower bound. The constructions were discovered through iterative interactions with a Large Language Model (LLM), illustrating the potential of LLMs for finding explicit combinatorial constructions.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_ca290788ab", "k": "Agentic Context Management: Solving Agent Memory and Cost by Treating Them as Lifecycle and Architecture Problems", "domain": "science", "knowledge": "Production AI agents' failures are less often due to an inability to reason well and more often because they cannot manage what is in their reasoning context: conversation histories, large prompts, large tool definitions, and ballooning tool outputs. Agents drown in their own accumulating history while paying a token cost that grows every turn, producing missing recalls within and across conversations. The incumbent response treats this as a storage-and-retrieval problem. We argue that framing is too narrow. Actively managing what an agent holds in mind is a lifecycle, not merely a store: it spans deciding what to remember, extracting and structuring it, choosing the right store per data type, consolidating and forgetting while preserving provenance, deciding what is relevant now, anticipating what is needed next, and compacting context to a budget without losing what matters. In serious production this operates not over a single user but across an organizational scope hierarchy. We name this discipline Agentic Context Management (ACM) and decompose it into five primitives: architecting, ingesting, scoping, anticipating, and compacting & consolidation. We then make the economic case: naive context accumulation grows token cost quadratically in conversation length, crude summarization buys linear cost at the price of an accuracy cliff, and only validated compaction achieves linear cost with preserved fidelity. We describe a reference implementation, Maximem Synap, that realizes the five primitives as a multi-tenant service and reports 92% on LongMemEval and 93.2% on LoCoMo under the configuration detailed in Section 6. We close with dimensions existing benchmarks do not yet capture, latency, token efficiency, and context-rot resistance, and the frontier of decision-level and organization-level context the category points toward.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_3c23f53a35", "k": "Artificial Epanorthosis: Why large language models overuse a classical rhetorical figure, and how to mitigate it", "domain": "science", "knowledge": "A rhetorical figure that Cicero and Quintilian catalogued two thousand years ago reappears, systematically, in the text of large language models: epanorthosis, the self-correction of the specimen «This is not a course. It is a journey of transformation». This essay argues that the overuse is a trained disposition, driven mainly by a training distribution rich in promotional prose and by preference tuning (RLHF) that rewards confident, emphatic phrasing; the left-to-right nature of generation is an amplifier rather than the root cause. Building on evidence that models diverge from human rhetorical style, and on Fontanier's classification of epanorthosis as a figure of thought, it sets out a programme that scores the figure against genre-specific human baselines through an Epanorthosis Index (density relative to the human rate). A first measurement, on three sizes of one instruction-tuned model family, finds mis-calibration by register in both directions: the models overshoot in oratory (about twofold, near threefold in Italian, concentrated in the larger tiers) and undershoot in informal question-and-answer writing, while matching humans in argument, journalism, and encyclopedic prose. Three constructive contributions follow: a survey of mitigation techniques centred on lightweight LoRA adapters; a demonstration, in Italian, that a one-line instruction cuts the figure by half to nearly three-quarters and that a supervised-fine-tuning adapter removes it almost entirely, with a scaling coefficient that dials the reduction back onto the human rate; and the argument that the target is calibration to the human rate for each genre, not elimination. It closes on the stakes: the real risk is that we begin to write like the machines.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_e038d96b1e", "k": "Toward Continuous Assurance for the Democratization of AI Agent Creation in Industry", "domain": "science", "knowledge": "AI agents are increasingly created inside organizations by non-engineering users through low-code, no-code, and conversational development environments. This democratization enables rapid local innovation, but it also creates a reliability gap: agents that appear to users as simple productivity artifacts may depend on changing models, tools, retrieval sources, permissions, prompts, schedules, and external services. These dependencies can cause silent degradation long after deployment, even when no user directly modifies the agent. This paper identifies the reliability challenge created by democratized AI agent creation and proposes a lightweight continuous-assurance framework for citizen-created organizational agents. The framework combines dependency mapping, readiness contracts, scheduled checks, diagnostics, and lifecycle governance to assess whether an agent remains operationally ready under expected conditions. We also present an initial prototype auditor and scenario-based assessment showing how the proposed taxonomy can be translated into practical checks and actionable remediation guidance.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_31a102b4d6", "k": "Compact Latent Coordination for Autonomous Vehicles at Unsignalized Intersections", "domain": "science", "knowledge": "Coordinating autonomous vehicles at unsignalized intersections remains a critical challenge for multi-agent reinforcement learning (MARL) systems, which typically struggle with combinatorial action spaces, reliance on privileged information, or rigid agent designs. We propose Master-Agent Proto-plan System (MAPS), a hierarchical deep reinforcement learning (DRL) architecture in which a centralized Master agent generates a compact, continuous embedding, denoted as proto-plan, that encodes a global coordination strategy. Decentralized Worker agents integrate this embedding with local observations to execute vehicle-specific control, decoupling strategic intent from tactical execution and enabling independent optimization of each module. As a proof-of-concept evaluation of this coordination mechanism, we test MAPS across 72 intersection configurations in HighwayEnv. MAPS achieves collision-free navigation while significantly reducing average travel time, outperforming state-of-the-art baselines. The learned proto-plans further exhibit robust generalization: a system trained with three agents achieves a 94% success rate when deployed zero-shot to five-agent scenarios, confirming that proto-plan-based hierarchical learning provides a promising framework for multi-vehicle coordination.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_47a7bd8b5d", "k": "Agentic coding without the cloud: evaluating open-weight large language models on longitudinal data preparation tasks", "domain": "science", "knowledge": "Large language models (LLMs) and agents are now widely used tools in code development, with data typically sent to third-party cloud-based models. Their adoption in research using personal data is constrained by governance requirements that typically prohibit data transmission to external services. Locally deployable open-weight models offer an alternative since sensitive data never leave the local environment. We introduce an open-source framework for evaluating the efficacy of AI agents powered by open-weight LLMs on one of the most persistent bottlenecks in research on longitudinal population studies: data preparation. The framework comprises: a curated ground-truth dataset (cleaning scripts preparing six sweeps of data from a British cohort study), task definitions encompassing tasks such as category harmonization and multi-wave merging, and automated routines for evaluating the LLM-produced R code and outputted data. We benchmark LLMs across the (consumer grade) deployment spectrum to assess their efficacy in 20 data preparation tasks (creation of 102 variables). Current state-of-the-art, 31-35B parameter models almost saturated our benchmark (\"average task completion\" up to 87.9%). The performance of open-weight LLMs running on consumer-grade hardware shows promise of a viable path toward AI-assisted data preparation in governance-restricted research settings. Our framework is publicly available at: https://github.com/UCL-ARC/RRBench.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_1701cbba98", "k": "Error Certificates for KV-Cache Eviction via Randomized Design", "domain": "science", "knowledge": "Deterministic KV-cache eviction keeps the top-$k$ tokens under an importance score and deletes the rest. We prove that this design cannot know what it destroyed: evicted values can be altered so that everything the serving system retains is unchanged while the true attention-output error grows arbitrarily, so no serving-time estimator of that error is consistent. Randomized eviction restores identifiability. With a Poisson-sampled tail at known inclusion probabilities, one logit offset performs the Hájek correction inside the softmax, and a survey-sampling variance estimator over the retained set becomes a per-step error certificate with 0.97 empirical coverage at no accuracy cost. On real workloads we pre-registered seven claims and lost three: question-aware eviction at 25--50\\% budgets is nearly free; output log-probability predicts failure better than the certificate; certificate-gated budget escalation adds nothing. What survives is attribution: the certificate separates cache-induced from inherent failures (AUC 0.73--0.75, against 0.47--0.54 for output confidence) and schedules recomputation better than random or confidence gating. Randomization buys attribution, not prediction.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_5fe38f9163", "k": "Thinkink: 2D Spatial Ink-native Interaction with LLMs", "domain": "science", "knowledge": "People often use handwritten notes and sketches to externalize ideas for ideation. To integrate large language models (LLMs) into this practice, we propose Thinkink. Prompts can be handwritten text or drawn sketches with LLM-generated responses visualized as ink-like text and sketches spatially integrated into a shared canvas. A semantic tree streamlines ink interpretation, and a lightweight UI provides explicit control using a state machine. The tool was designed using a three-stage process. A formative study (N=12) examined current practices with conventional and digital inking methods. The results informed a technical probe for a diagnostic study (N=6) identifying usability and human-LLM interaction challenges. This motivated the design of Thinkink, with a final study (N=10) examining how people incorporate it into their ideation practices. We contribute design implications and a tool for ink-native LLM interaction where the user and LLM write and draw in a shared 2D canvas.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_1084ef42c2", "k": "AREX: Towards a Recursively Self-Improving Agent for Deep Research", "domain": "science", "knowledge": "Deep research requires agents to find answers that jointly satisfy multiple constraints. Discovering such answers is costly, whereas verifying a candidate can often be decomposed into tractable constraint-wise checks. This discovery--verification asymmetry suggests that a research agent should do more than simply search longer: it should recursively improve its current answer by verifying intermediate results and using the partially verified state to guide subsequent refinement. We introduce AREX, a family of Recursively Self-Improving (RSI) deep research agents. AREX alternates between an inner research loop that gathers evidence and constructs a provisional answer, and an outer self-improvement loop that audits the answer constraint-wise, identifies unresolved claims, and launches targeted follow-up research. To sustain RSI over long horizons, AREX learns an autonomous context-update tool that compresses growing interaction history into a compact improvement state preserving verified evidence and unresolved constraints, without relying on an external model. We train AREX on verified synthetic tasks and high-quality trajectories through agentic mid-training and long-horizon reinforcement learning. To mitigate sparse final rewards during long horizon learning, we emphasize key steps where decisive evidence is acquired or erroneous research directions are corrected. We instantiate a dense 4B model and a 122B-A10B Mixture-of-Experts model. Across BrowseComp, WideSearch, DeepSearchQA, Humanity's Last Exam (HLE), and other reasoning and tool-use benchmarks, AREX substantially outperforms comparable-scale baselines and remains competitive with models using substantially more activated parameters.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_59fc576eaa", "k": "Detecting LLM-Generated Tokens in Human--LLM Coauthored Text", "domain": "science", "knowledge": "The rise of human-AI collaborative writing has created a growing need for fine-grained detection methods that support localizing likely LLM-generated content in mixed-authorship documents. Existing methods for detecting LLM-generated text mainly focus on document-level classification and cannot identify which parts of the text are generated by LLMs. This paper introduces a new method to address this urgent need. Our method operates at the token level, the natural unit of modern language models, and builds on existing token-level detection scores. The key idea is to smooth adjacent token scores to reduce their variability, while using an adaptive Lepski-type rule to select the bandwidth according to the local authorship structure. Our method is simple to implement and does not require token-level labeled data for training. Theoretically, we characterize this trade-off and show that the proposed method achieves favorable mean square error performance in estimating the underlying signal. Empirically, we demonstrate strong performance of our method against a wide range of baselines in both synthetic datasets and a realistic dataset. We deploy a publicly accessible website that implements the methods as well.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_5a3e631260", "k": "RUMBA: Russian User Memory Benchmark", "domain": "science", "knowledge": "The ability to handle long-term memory in LLMs is becoming increasingly critical, yet existing benchmarks remain English-centric and rely on aggregate retrieval metrics, failing to capture interactions between long-range context, temporal information, and reasoning. To address this, we introduce RUMBA (Russian User Memory BenchmArk) - a new benchmark for long-term conversational memory that provides a fine-grained taxonomy of memory-centric question types and a unified methodology accounting for semantic type, session scope, temporal reasoning, and the explicitness of temporal expressions. RUMBA consists of timestamped user-assistant dialogues with QA pairs requiring retrieval, combination, and reasoning across sessions. While designed for Russian, we also provide an aligned English subset under the same methodology. We evaluate contemporary memory systems and long-context models, and show how RUMBA serves as a diagnostic tool to analyze model behavior across benchmark slices and identify strengths and failure modes of different memory mechanisms.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_85470a283a", "k": "Agent-Guided Relational Concept Discovery: Toward Interpretable Surgical Margin Assessment", "domain": "science", "knowledge": "Deep learning models can effectively use Rapid Evaporative Ionization Mass Spectrometry (REIMS) data for surgical margin assessment. However, their clinical adoption remains challenging due to limited generalization to operating room conditions. This difficulty arises because models are typically trained on labeled spectra collected from resected tissue samples, while they must operate on noisy, unlabeled data acquired directly during surgery. In addition, the black-box nature of deep learning models makes it difficult to understand and systematically improve their behavior. Concept-based learning offers a promising way to address these challenges by mapping raw measurements to human-understandable concepts. However, supervised concept-based approaches rely on concept annotations, which are difficult to obtain in complex mass spectrometry workflows. We propose Agent-Guided Concept Discovery, a framework that learns meaningful concepts directly from data without requiring predefined concept labels. During training, a reasoning agent refines semantic descriptions of the learned concepts and adaptively adjusts their weight based on diagnostic relevance. These concepts are further grounded using a biochemical knowledge graph to ensure consistency with known metabolic relationships. Across Skin and Breast Cancer datasets, our model improves balanced accuracy and sensitivity over the baseline. In a representative intraoperative case, it shows fewer false positives, indicating better generalization to surgical conditions.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_3db56f0feb", "k": "Adaptive Identity Anchoring: Closed-Loop Keyframe Placement for Synthetic Paired Supervision in Video Face Swapping", "domain": "science", "knowledge": "Video face swapping has no natural paired supervision: no real footage exists of one person's face performing another person's video. The strongest current answer, DreamID-V's SyncID-Pipe, mints pairs by replacing the identity in exactly two frames of a real clip -- the first and the last -- and regenerating the rest from a pose sequence alone. Pose carries no appearance evidence of the swapped-in identity, so over long clips, occlusions, and extreme pose excursions the synthesized identity has a long unanchored span on which to drift; no published ablation examines anchor count or placement. We propose Adaptive Identity Anchoring (AIA): (i) generalize the synthesizer to arbitrary anchor sets, architecturally natural for diffusion-forcing-style transformers where conditioning on a frame is clamping its tokens to zero noise; (ii) place anchors by a closed feedback loop that scores every generated frame against the real reference identity and inserts an image-face-swapped anchor at the worst-scoring frame until the pair passes a threshold or exhausts a budget; (iii) reuse the loop's verdict as an automatic data filter. A second pathology, the beauty-filter look of over-smoothed skin, has the same root cause: micro-texture, like identity, is priced by none of the pipeline's objectives. We therefore pair AIA with Reality-Referenced Texture Restoration: matched re-graining from each real frame's non-face regions, band-split transfer of sub-identity micro-texture from the real footage, and a second, spectral acceptance channel refereed by the footage's own spectrum. Identity-anchor density, we argue, is a controllable quality dial, and we specify falsifiable experiments -- drift-versus-gap curves, uniform-versus-adaptive placement at matched budgets, student training on AIA-minted data, and texture ablations with a human beauty-filter study -- that would validate or refute the proposal.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_a4c22a60a1", "k": "Token Budget Saturation and Mechanistic Early Detection of Reasoning Non-Convergence in Chain-of-Thought Models", "domain": "science", "knowledge": "Chain-of-thought reasoning models such as DeepSeek-R1-Distill-Qwen-7B exhibit a bimodal convergence pattern: generations either terminate within a token budget (converged) or exhaust it without reaching a conclusion (non-converged). We characterize this phenomenon empirically, showing that converged generations achieve 90.3% accuracy on AIME 1983-2024 while non-converged ones achieve only 6.6%, with an overall convergence rate of 62.0%. We then ask whether this outcome is detectable early in the thinking chain using internal model representations. Training linear probes on hidden-state activations at token positions 50-300, we find that layer-20 activations at token 150 achieve AUC 0.608 (+-0.080, 5-fold CV), reliably above chance even at token 50. Activation probes consistently outperform behavioral baselines derived from token entropy and repetition statistics. A sweep-level permutation test yields p=0.063 (100,000 permutations), consistent with a modest signal that our sample size cannot confirm at conventional thresholds. These findings suggest that convergence fate is partially encoded in intermediate representations well before the generation ends, opening a path toward early-exit inference and adaptive compute allocation.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_fe35af072a", "k": "Bridging the Gap Between Plausibility and Admissibility: Constraint-Aware Flow Maps for Dynamic Graph Systems", "domain": "science", "knowledge": "Generative models can support decision-making under uncertainty by producing ensembles of plausible future system trajectories, but statistical plausibility does not ensure structural feasibility. This study investigates whether post-sampling symbolic constraints can improve the reliability of generative trajectory modeling in dynamic graph-structured systems. A conditional diffusion model generates future graph-state trajectories from partial observations, while an external symbolic layer applies hard filtering, soft weighting, or projection-based repair. The framework is evaluated on two controlled synthetic regimes: a compact graph and a medium-complexity dependency graph, using metrics for structural validity, sample efficiency, diversity, robustness, and calibration. In the compact regime, the model produces an invalid probability mass of 0.002996, indicating an almost entirely admissible trajectory manifold. Under the same architecture and training protocol, invalid mass increases to 0.155929 in the medium-complexity regime. Hard filtering removes all invalid retained trajectories while preserving 84.4% of generated samples, whereas soft weighting preserves effective sample size but yields only limited validity gains. Family-level analysis shows that dependency constraints account for nearly all observed inadmissibility. These results indicate that statistical plausibility and structural admissibility are distinct reliability properties and that symbolic constraint handling becomes more valuable as graph-structural complexity increases.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_268cc4b6b7", "k": "PATS: Policy-Aware Training Scaffolding for Agentic Reinforcement Learning", "domain": "science", "knowledge": "In long-horizon LLM agent reinforcement learning, weak policies often repeat similar failures, producing uninformative rollout trajectories and limiting effective policy optimization. Existing skill-centric methods improve exploration by optimizing, filtering, or internalizing reusable skills. However, they remain centered on the skills themselves rather than being designed as adaptive training-time support for the evolving policy. To address this, we propose a policy-centric training paradigm that reframes skills as a dynamic training scaffold. Our framework, Pats, converts rollout groups from the latest policy into evidence cards and uses task-specific evaluation to adjust the context used in subsequent rollouts. Concrete guidance helps weak policies to complete challenging tasks. As policy improves, redundant context is revised or removed to reduce reliance on explicit guidance while preserving useful rollout variation. The policy is optimized with environmental rewards using standard RLVR, and the training scaffold is discarded at deployment. On ALFWorld and WebShop, Pats improves over strong baselines by up to 18.6%. Across seven search-augmented QA benchmarks, it remains competitive while using 32.1% fewer prompt tokens than the baseline.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_cbe38d0579", "k": "Logical Regression for Planning with Axioms", "domain": "science", "knowledge": "In automated planning, logical regression is an operation that returns the most general condition necessary for an action to achieve a particular formula. It has many applications, such as allowing for more robust plan execution and providing compact policies for non-deterministic planning. Although relatively simple to calculate in basic planning settings, logical regression becomes significantly more complex when additional factors, such as axioms, are present. We introduce a methodology for approximating the logical regression of an action in a domain that includes axioms; an approximation that limits conditions to partial states. Our method produces minimal partial states while avoiding the recalculation of axioms. To demonstrate the impact of our methods, we embed our form of regression in an execution monitoring context, a well-established setting that can benefit greatly from logical regression. Our results show that this form of regression can dramatically generalize partial states across multiple domains, reducing the number of variables considered for execution monitoring by up to 70%, and demonstrate that the resulting execution monitor is robust enough to recover frequently in an environment with unexpected changes: several domains recover over 50% of the time in our tests.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_a5d902b16c", "k": "Euclid-MCP: A Model Context Protocol Server for Deterministic Logical Reasoning via Prolog", "domain": "science", "knowledge": "Large Language Models (LLMs) excel at natural language understanding and generation but remain unreliable for multi-step logical reasoning, especially in safety-critical or compliance-sensitive domains. Recent neuro-symbolic approaches address this gap by coupling neural models with external symbolic engines, yet most integrations are bespoke and lack a standardized interface for tool-augmented agents. This paper presents Euclid-MCP, an open-source MCP server that provides deterministic logical reasoning via SWI-Prolog. Euclid-MCP introduces Euclid-IR, an engine-agnostic intermediate representation for Horn-clause logic that is human-readable, easy for LLMs to generate, and straightforward to compile into Prolog or alternative backends. The server exposes a compact tool interface that supports a translate-run-inspect-repair loop, enabling LLM clients to delegate inference while retaining full access to proof traces and derivation logs. We evaluate Euclid-MCP on a realistic IT security and compliance use case. Results show that while LLMs alone are sufficient on small knowledge bases, they hallucinate systematically on larger problems, whereas Euclid-MCP delivers exact answers with lower latency and more compact outputs. We argue that semantic RAG is fundamentally unsuited for rule enforcement, and that Euclid-MCP can serve as a stable, shared reasoning substrate for both RAG-based assistants and agentic systems.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_c38a75e69b", "k": "Cycle-Consistent and Uncertainty-Aware Neural Surrogates for Tokamak Edge Plasmas", "domain": "science", "knowledge": "The boundary and divertor plasma govern how a tokamak exhausts power and particles, setting heat fluxes, target conditions, and the onset of detachment. Predicting these quantities is essential for operating current and future devices, but edge simulations that resolve them are too slow for parameter scans, optimization, or real-time control. Machine-learning surrogates offer a fast alternative, yet most are forward-only: they cannot recover input parameters from observations or assess the reliability of their predictions. We introduce a cycle-consistent neural surrogate for edge plasmas, combining a conditional U-Net forward model with an optimization-based inverse method built on the frozen forward network. The forward model maps five control parameters to two-dimensional plasma-state fields on the SOLPS-ITER mesh; the inverse method enforces consistency between forward and inverse predictions, a self-supervised quality check needing no ground-truth labels at inference. An ensemble of multilayer perceptrons also predicts electron temperature and density profiles at the outboard midplane and divertor targets, with uncertainty estimates that flag where more simulations are needed. The forward model achieves normalized root-mean-square errors below 2.6% and Pearson correlations above 0.95 for all fields. Cycle-consistency regularization raises the average cyclical $R^2$ from 0.59 to 0.99 without degrading forward accuracy and enables recovery of the core fueling rate; all five control parameters are recovered with Pearson $r\\ge0.97$. A $k$-d tree warm start yields a database completion rate above 95%, versus roughly 30% outright failures when cold-started. With about $4\\times10^6$ parameters, the model produces full 2D predictions in milliseconds, five to six orders of magnitude faster than SOLPS-ITER, enabling real-time control, parameter scans, uncertainty analysis, and digital twins.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_4d0cefcb14", "k": "MSBraM: A Multi-scale Self-supervised Brain Foundation Model for Hierarchical EEG Dynamics Learning", "domain": "science", "knowledge": "Self-supervised foundation models have recently shown strong potential for electroencephalogram (EEG)-based analysis. However, existing approaches struggle to capture the inherently multi-scale temporal structure of EEG signals, where local neural patterns and long-range dependencies jointly encode task-relevant information. This limitation hampers cross-scale representation learning and generalization across diverse downstream tasks. To address this challenge, we propose MSBraM, a Multi-Scale self-supervised Brain foundation Model designed to learn hierarchical EEG representations. MSBraM follows a two-stage pretraining framework. First, a multi-scale neural tokenizer discretizes raw EEG signals into semantic codes at different temporal resolutions via vector-quantized reconstruction. Second, the model is pretrained to predict masked codes using a curriculum multi-scale masking strategy, progressively integrating fine-grained local patterns with global temporal context. We pretrain MSBraM on over 2,400 hours of EEG data and evaluate it across 10 downstream tasks on 12 public datasets. Extensive experiments show that MSBraM achieves superior performance on other state-of-the-art pretrained models, demonstrating strong generalization and transferability. These results indicate that explicitly modeling multi-scale temporal dynamics is critical for effective EEG foundation models.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_1681849f07", "k": "When Are Reasoning-Based Guardrails Not Efficient? ResponseGuard: A Fast Vision-Language Guard for Real-Time Moderation", "domain": "science", "knowledge": "A vision-language AI assistant returns its answer as a stream of generated tokens. Therefore, a safety guard that watches that answer has to keep up with the stream and stop a harmful reply before a user reads it. Recent vision-language guardrails instead generate a chain of thought before they issue a verdict. They believe that step-by-step reasoning yields a safer guard. This design makes the guard heavy and slow, since the model must decode many tokens for harmfulness detection. We pose the question of whether a vision-language guard really needs to reason in order to screen a response. We answer with a guard that has no chain. ResponseGuard reads a harmful verdict from a single pooled representation of the request, the response, and the image in one forward pass. Across a standard multimodal guardrail benchmark, our 2B ResponseGuard outperforms a recent 3B reasoning-based vision-language guard on response harmfulness detection, without any reasoning and at about 150 times lower time cost. On request harmfulness the reasoning guard retains an overall lead, and the remaining gap on both tracks sits on the image-only cells. We observe that the gap may stem from the frozen vision encoders that both designs use rather than from the missing chain. We have also found the reasoning guard directs almost none of its verdict attention to the image. Based on a single-pass detection, ResponseGuard can screen an answer sentence by sentence as it streams and stop a harmful answer before it finishes. For guarding the response of a vision-language model, a calibrated single-pass label may provide a sufficient safety signal. We fully release all source code, trained models, and datasets at https://github.com/ndb796/ResponseGuard.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_0477356763", "k": "VoLN: Vision-Only Long-Horizon Navigation---Paradigm, Benchmark, and Method", "domain": "science", "knowledge": "Vision-and-Language Navigation (VLN) enables embodied agents to follow natural-language instructions. However, route-level instructions commonly encode spatial priors, such as orientation, distance, and layout, that are not explicitly available from onboard sensing at deployment in open, GPS-denied environments. Benchmark performance under such interfaces therefore jointly reflects visual navigation ability and the use of route structure explicitly supplied by the task description. As a complementary formulation, we propose Vision-Only Long-Horizon Navigation (VoLN), which shifts route-relevant information from externally supplied instructions and global guidance to locally observable in-scene cues. In VoLN, goal views specify the destination, while route-relevant information is available only through locally observable in-scene cues that the agent must detect, interpret, and select online. We instantiate VoLN for aerial navigation through VoLN-UAV, a 7,210-episode benchmark that combines long-horizon goal-directed flight, continuous 3D motion, large viewpoint changes, and context-dependent beacon selection. We further provide VoLN-MLLM as an initial reference baseline. It aligns self-supervised visual features with a structured semantic space and predicts short-horizon waypoint segments from observation history, goal views, retrieved visual--semantic tokens, and proprioception. On the five-environment Test-Unseen split, it obtains success rates of 7.4%, 4.5%, and 1.8% on Easy, Normal, and Hard episodes, respectively. These results provide an initial evaluation of VoLN and reveal substantial remaining challenges in long-horizon evidence integration, cross-view goal matching, and closed-loop stability. Project page: https://admire-ljb.github.io/VoLN-UAV/", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_e6edc6b3f7", "k": "Multimodal Pretraining for Generalizable EEG Representation Learning", "domain": "science", "knowledge": "Electroencephalography (EEG) models used for epilepsy are often limited to specific datasets and tasks. This limited approach can make it challenging to apply these models across different datasets or in various situations. However, recent studies in foundation models and self-supervised learning suggest that an adaptable EEG backbone could support a range of EEG related tasks. In this study, we have developed a multimodal EEG foundation model that combines a raw signal encoder based on the Mamba architecture, a Vision Transformer (ViT)-style encoder for time-frequency data, and a lightweight encoder for text, all within a shared embedding space. The pretraining process relies on several innovative techniques, such as masked modeling, cross-view contrastive alignment, and temporal consistency losses. These methods are designed to create rich, seizure-relevant representations without requiring labeled data. To assess the efficacy and generalization of our pretrained model, we fine-tuned it on the canonical CHB-MIT seizure detection benchmark and additional seizure detection datasets, and conducted extensive experiments comparing different model variants. On the standard CHB-MIT split, our best single model achieved an AUROC of 0.874, and an ensemble variant reached 0.878 AUROC, representing state-of-the-art performance on this benchmark. In addition to standard train-test splits, we evaluated performance under a leave-one-subject-out (LOSO) protocol, which is rarely reported in prior EEG seizure modeling work and highlights the difficulty of patient-independent seizure detection, with a mean LOSO balanced accuracy of 0.558 across 19 subjects. Across datasets and evaluation settings, our multimodal foundation model enabled robust seizure detection and straightforward adaptation to new seizure detection scenarios, while also supporting interpretable seizure localization.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_efe2074d8e", "k": "Towards Faithful Graph Explanations with Synergistic Edge Effects via Granular Balls", "domain": "science", "knowledge": "Instance-level explanations aim to reveal the rationale behind a model's decisions for a specific graph. Previous methods explain graph neural networks (GNNs) by selecting important edges to induce subgraphs, where edge importance is assessed by perturbing each edge and observing changes in the model predictions. However, they often neglect the synergistic effects among edges, which are crucial for accurately characterizing edge importance. To address this issue, we propose SeeExplainer, a parameter-free explainer to interpret GNNs. Specifically, we first introduce a granular-ball graph refinement mechanism that decomposes a graph into several disjoint granular-balls with no fixed size, and utilize them as nodes to construct a structural graph. This process can better capture the synergistic effects among edges. Then, we perturb nodes and edges in the structural graph to generate explanatory subgraphs based on their respective contributions. Experiments on several graph classification datasets of different networks show that SeeExplainer outperforms state-of-the-art baselines.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_f872602032", "k": "Mean-to-Score Discrete Diffusion: Posterior-Mean Denoisers for Score Entropy", "domain": "science", "knowledge": "Score Entropy Discrete Diffusion (SEDD) parameterizes discrete reverse processes with unconstrained positive score ratios. While positivity guarantees nonnegative reverse jump rates, it does not ensure Bayes realizability: ratios at a noisy state need not be jointly induced by any clean-token posterior under the forward kernel. The score-entropy loss has the correct population optimum but does not enforce this constraint away from it. In a trained pure-uniform SEDD checkpoint, roughly one quarter of complete score vectors violate the coordinate box, while more than half lie inside it yet remain materially incompatible with any valid posterior. Such violations can produce negative pre-normalization weights in finite-step sampling. Projecting raw scores onto the bridge polytope removes all observed negative weights and improves external generative PPL from $203.6$ to $175.1$ without changing the sampler. We introduce \\emph{mean-to-score} (M2S), which predicts a clean-token posterior mean and converts it to the score through an exact kernel-dependent linear map. The construction applies to any known coordinate-wise continuous-time Markov chain (CTMC) satisfying a mild support condition. For uniform corruption, it maps the probability simplex onto the bridge polytope; for absorbing-mask corruption, the resulting objective recovers MD4 exactly. In a controlled 28.4M-parameter CIFAR-10 comparison, M2S lowers test BPD from $3.173$ to $3.129$ and FID-50k from $\\CifarSEDDFID$ to $\\CifarMtwoSFID$. A 170M-parameter M2S model trained on about 262B OpenWebText token slots outperforms the evaluated pure-uniform SEDD, GIDD, and Neural CTMC checkpoints at every tested sampling budget, reaching generative PPL $143.3$ at 128 steps versus $183.6$ for the strongest pure-uniform baseline.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_d8990dfd80", "k": "DINOde: Continuous Vision-Text Alignment for Open-Vocabulary Semantic Segmentation", "domain": "science", "knowledge": "Open-vocabulary semantic segmentation (OVSS) leverages textual semantics to segment objects beyond predefined categories. While the self-supervised model DINOv3 provides strong structured visual representations, its lack of native textual alignment hinders its direct application to OVSS. To bridge this gap, we propose DINOde, an ODE-based framework that continuously aligns CLIP text embeddings with the DINO visual manifold. Our approach employs two complementary components: (i) Semantic Text Flow (STF), which evolves text embeddings toward the DINO manifold through a continuous ODE trajectory, and (ii) Global Context Flow (GCF), which progressively refines the holistic image representation carried by DINO's CLS token. To preserve the hyperspherical geometry of the feature space during this evolution, we further introduce Velocity Tangent Projection, which constrains the learned velocity field to the tangent space. By modeling alignment as a continuous trajectory, DINOde avoids the manifold entanglement inherent in discrete MLP projections and yields more robust cross-modal alignment. Extensive experiments demonstrate that DINOde consistently outperforms existing methods and achieves state-of-the-art performance across multiple OVSS benchmarks. The code is available at https://github.com/yoon307/DINOde.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_016f1f1dba", "k": "Hilbert Operator for Progressive Encoding (HOPE): A Mathematical Framework for Deconstructing Learned Representations in Deep Networks", "domain": "science", "knowledge": "Deep neural networks encode complex representations, but deconstructing this internal knowledge remains a challenge. Given the link between learning and compression, network compression offers a promising lens to analyze this knowledge. However, standard compression heuristics often suffer from scale symmetries and architectural biases. To resolve these, we introduce Hilbert Operator for Progressive Encoding (HOPE), a mathematical framework to gradually deconstruct the representations in trained network weights. HOPE shifts network compression from the discrete domain into a Hilbert space of continuous functions. By modeling individual neurons as rank-1 Hilbert-Schmidt operators, HOPE unifies pruning and neuron merging as low-rank subspace projection. Extending this formulation, HOPE introduces macro block eviction to encompass multi-layer structures like entire residual pathways under the same unified metric. This unified approach enables unbiased architectural decisions across layers with different types and sizes. HOPE is a data-free and hyperparameter-free framework. We present proof-of-concept experiments in model compression and fine-tuning to highlight the practical potential of our theory.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_685c55a32a", "k": "SPORD: A Simulation-Propose-then-OR-Dispose Approach for Supply Chain Planning", "domain": "science", "knowledge": "For years, supply chain planning at e-commerce firms has operated as a collection of isolated projects. Each planning task from static network planning to dynamic warehouse assortment planning requires analysts to spend weeks building models from scratch, calibrating and persuading executives to act on outputs they cannot verify. Three barriers drive this: bespoke models proliferate because standardization is difficult (operational fragmentation); once unified, the combinatorial scale of millions of SKUs, thousands of nodes, and intricate routing logic exceeds what solvers can handle within a tight window (computational intractability); and a mathematically optimal solution still fails to be implemented if the executives do not trust it (implementation hurdle). To bridge this gap, we propose and implement the Simulation-Propose-then-OR-Dispose method, deployed as JD.com's NetSim platform. The central insight is decoupling: simulation proposes by generating and evaluating the full set of operationally valid candidate paths-absorbing all idiosyncratic business logic, while an integer program disposes by selecting the globally optimal subset. Computationally, matrix-vectorized CPU/GPU accelerated simulation achieves a 10-100 times speedup over serial methods, and a list scheduling algorithm reduces coupled-order processing from hours to minutes. Operationally, we establish a closed loop via an intelligent diagnosis engine. Since 2025, NetSim has optimized end to-end services for over 20,000 suppliers, the cross-regional fulfillment rate dropped from 6.1% to 4.9%, and the average monthly carbon reduction is approximately 5,745 tCO2e. SPORD moves simulation from monitoring to active planning. The transparent outputs turn skeptical executives into engaged collaborators, and the modular architecture ensures that the next planning requires just configuration, not reconstruction.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_7cb71f0e98", "k": "Regulating autonomous and agentic AI", "domain": "science", "knowledge": "Regulating activities where regulatees use autonomous and agentic AI is challenging. Regulatory assumptions about regulatee knowledge and control no longer hold true; much of that lies elsewhere in the AI supply chain which thus needs to be brought within the scope of regulation. Governance systems for autonomous AI cannot replicate existing governance models, but need a fresh approach. Retrospective supervisory oversight becomes ineffective as a risk management tool, and AI autonomy generates new systemic risks which require new solutions. This paper investigate four regulatory systems: UK regulation of content platforms, data protection, UK financial services, and the EU AI Act\\'92s cross-sectoral regime. It analyses the challenges posed by autonomous and agentic AI and proposes potential solutions which regulators might adopt. These will transform regulation from a reactive process to an active one, and assist it in adapting to the challenges of AI autonomy.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_b4c692c546", "k": "M$^3$-Gen: Interpretable Multimodal Generation of Gene Expression Profiles Using Clinical and Imaging Data", "domain": "science", "knowledge": "Integrating heterogeneous biomedical data, including clinical metadata, histopathology images, and molecular profiles, is crucial for comprehensive disease understanding. However, gene expression data acquisition remains constrained by high costs and privacy concerns, limiting its use in multimodal research and AI-driven applications. We present MultiModal Molecular Generation (M$^3$-Gen), a novel framework for the generation of gene expression profiles by conditioning a Generative Adversarial Network on histopathology images and clinical metadata. M$^3$-Gen learns a unified latent representation from the clinical variables and the images, leveraging contrastive learning, and exploits the embeddings of the two modalities to guide a generative model in producing biologically coherent gene expression profiles. Evaluations on the TCGA dataset demonstrate that M$^3$-Gen generates realistic and functionally meaningful gene expression data. Importantly, by integrating multiple modalities in an attention-based mechanism, M$^3$-Gen provides intrinsic explainability: it allows the identification of which regions of the histopathology images most strongly influenced the generation of specific gene expression profiles, making the model's decisions interpretable by design.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_9dca05e9ad", "k": "Phonetic forced alignment for low-resource language varieties: Model training and evaluation on Chengdu Mandarin", "domain": "science", "knowledge": "Phonetic forced alignment is a key technique in phonetic research, yet existing alignment systems lack specialized models for low-resource language varieties. We address this by training text-dependent and text-independent aligners for Chengdu Mandarin using a 17-hour corpus and a custom G2P dictionary. We trained a text-dependent GMM-HMM model (Chengdu-MFA) and fine-tuned a pretrained audio encoder on frame classification with Chengdu-MFA's pseudo label for text-independent alignment (Chengdu-FC). Evaluation on an expert-annotated test set show that both methods significantly outperform Standard Mandarin baselines. Chengdu-MFA reduced average phone boundary differences by 31.8%, while Chengdu-FC achieved a 61.2% reduction. This work establishes a practical bootstrapping pipeline for developing accurate aligners for under-resourced varieties without labor- and time-intensive manual annotation.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_6217c7d105", "k": "From Static Bibliometrics to Dynamic Knowledge Graphs: An LLM-Powered Framework for Modernizing Science, Technology, and Innovation (STI) Analytics", "domain": "science", "knowledge": "Bibliometric indicators - citation counts, h-indexes, co-authorship networks - have long anchored science, technology, and innovation (STI) analytics, yet suffer from temporal lag, semantic shallowness, and an inability to capture the non-linear dynamics of contemporary knowledge ecosystems. Dynamic knowledge graphs and large language models (LLMs) have each been proposed as remedies, but neither is sufficient alone: existing scholarly knowledge graphs remain largely static, while LLM-driven pipelines are prone to hallucination, opacity, and corpus bias without structured grounding. This paper proposes a hybrid, symbolic-first framework integrating all three traditions under explicit methodological constraint. Organized across five layers - an open scholarly data backbone, a dynamic versioned knowledge graph, a constrained LLM-assisted semantic augmentation layer, a multi-layer validation pipeline, and an analytics layer - the framework positions LLMs strictly as generators of provisional candidate enrichments. Candidates become analytically admissible only after passing structural, evidentiary, comparative, and selective expert validation, with full provenance recorded at every stage. The analytics layer supports both established bibliometric indicators and extended graph-based analyses, including trend emergence detection, science-to-technology pathway mapping, and policy-oriented gap analysis. The framework's central theoretical contribution is treating validation as the mediating principle between semantic flexibility and epistemic discipline, enabling STI analytics that is semantically richer and temporally more responsive than static bibliometrics while remaining aligned with the evidentiary standards of science-of-science research. Governance considerations addressing reproducibility, bias, and auditability are also discussed.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_02d882080c", "k": "Toward cryptographically verifiable authorization for autonomous AI agents: A security hypothesis, preliminary formal model, and proof-of-concept implementation", "domain": "science", "knowledge": "Autonomous AI agents increasingly execute actions, invoke tools, and operate on protected resources with limited human oversight. Existing authentication and authorization mechanisms establish identity and delegate authority, but do not inherently provide cryptographic evidence that a concrete request issued by a specific agent satisfies the applicable policy in a specific execution context. This paper hypothesizes that agent authorization can be formalized as a cryptographically verifiable relation, denoted $R_{CVA}$, that jointly binds an agent principal, a concrete authorization request, an execution context, and the satisfaction of an applicable policy, while selectively preserving the confidentiality of private authorization attributes. We introduce a preliminary formal abstraction for Cryptographically Verifiable Agent Authorization (CVA), define a compact set of candidate security properties including authorization soundness, principal binding, request binding, policy binding, and replay resistance, and provide an executable zero-knowledge proof of concept that instantiates selected elements of the model over a Groth16 zk-SNARK construction. We further identify and formalize the structural separation among identity binding, authorization-request binding, and runtime execution binding as a central open problem in the design of secure agentic systems (a distinction {not explicitly addressed by} current agentic security frameworks) and present a falsifiable research agenda for its resolution.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_057dd98fc5", "k": "GRADRAG: Cross-Component Prompt Adaptation for Coordinated Multi-Agent RAG", "domain": "science", "knowledge": "Retrieval-Augmented Generation (RAG) systems increasingly employ multiple LLM agents. Yet, most prior work optimizes components in isolation rather than coordinating improvements across the pipeline. We introduce GRADRAG, a framework for cross-component prompt adaptation that models the RAG pipeline as a computational graph and propagates structured evaluation feedback to update upstream agents. An Evaluator critiques downstream answers and supporting evidence, producing actionable feedback that a Prompt Optimizer uses to iteratively update adaptive agents, such as retrievers, graph constructors, and answerers. The Evaluator also triggers early stopping when the output is deemed satisfactory. We evaluate GRADRAG on the SQUALITY and QMSUM benchmarks under two retrieval paradigms: flat chunk-based retrieval using IRCoT-style query refinement (Trivedi et al., 2023), and graph-based retrieval that constructs and iteratively enriches an entity-relation graph from the document. Across both settings, GRADRAG consistently outperforms one-step refinement baselines that update only the final generator, achieving a 12-15 percentage point net preference margin in LLM-judged pairwise comparisons, with most gains realized within two refinement iterations.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_cf7a2a0364", "k": "PC-Edit: Prompt-Contrastive Region Discovery and Region-Guided Editing", "domain": "science", "knowledge": "Replacing an object with one that differs in category or shape requires complete source removal, natural target formation unconstrained by the source silhouette, and preservation of unrelated content. Existing training-free editors either localize edits from terminal predictions under source and target prompts or preserve unrelated content through spatially unselective source-feature reuse without explicit region discovery. Before reaching the terminal predictions, prompt-induced semantic differences undergo additional network transformations that may obscure their spatial localization, reducing localization precision. Spatially unselective feature reuse forces a trade-off between edit completeness and background preservation. Therefore, we propose PC-Edit, a prompt-contrastive framework for training-free MM-DiT editing. PC-Edit contrasts the image-token attention outputs under the source and target prompts, capturing prompt-induced semantic differences directly where text-conditioned information is delivered to image tokens. The same contrast identifies a source-erasure region during inversion and a target-emergence region during denoising. Their union suppresses source remnants while allowing the target object to form naturally. PC-Edit further couples region discovery and background preservation within each sampling step by estimating the current edit region from preceding attention blocks and immediately injecting cached source K/V features outside it in subsequent blocks, thereby protecting unrelated content before the latent update. Experiments on PIE-Bench and our EditRegion-Bench, with human-verified edit-region annotations for single- and multi-object addition and replacement, show that PC-Edit achieves the best editing quality and background preservation among methods without user-specified edit regions.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_96f884babb", "k": "Scaling Up Formal Representation of Clinical Trial Protocols in Ensemble Logic Using LLMs: A Preliminary Study", "domain": "science", "knowledge": "The reliance on unstructured free text for documenting clinical trial protocols creates a significant barrier to automated reasoning, cohort discovery, and trial simulation. The lack of formal structure obscures critical temporal phenotypes, such as dynamic eligibility criteria and event timing constraints. Although Temporal Ensemble Logic (TEL) offers an expressive framework for modeling these elements, manual encoding remains a prohibitive bottleneck. We introduce the CT-TEL workflow: a scalable pipeline leveraging Large Language Models (LLMs) to translate narrative clinical protocols into TEL formulas. We applied CT-TEL to generate logical models for 23 real-world trials from ClinicalTrials.gov. We evaluated translation fidelity via a back-translation approach, using LLMs to convert TEL formulas back into natural language and measuring semantic similarity against source texts. The resulting semantic retention suggests that LLMs may offer a pathway for mapping informal protocols to computable logic, providing preliminary evidence toward scalable clinical trial emulation within the emerging \"Symbolic Biomedicine\" paradigm championed by the corresponding author.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_13f5ffb426", "k": "AI Assistants Overassist", "domain": "science", "knowledge": "Large language models (LLMs) are increasingly used as tutors and thought partners, helping users reason through problems. While guidance from AI assistants can scaffold thinking and foster learning, such benefits depend on how they help--for instance, intervening too early or too frequently may hinder true learning and cognitive engagement. Yet how AI systems navigate intervention decisions during problem-solving remains poorly understood. Here, we introduce Int-Bench, a simulation-based benchmark for evaluating LLM interventions during learning. Int-Bench simulates a \"student\" solving a problem while a \"teacher\" monitors the student's reasoning and decides whether, when, and how to intervene. Across three domains--code debugging, mathematics, and brain teasers--we evaluate LLM teachers on the frequency and timing of interventions, as well as their impact on both immediate task success and generalization to new problems. We also compare LLMs to humans, finding that LLMs intervene more frequently and earlier than humans. Moreover, in contrast to humans, they tend to provide complete solutions rather than targeted hints. These findings suggest that current LLM assistants often optimize for short-term success rather than supporting the reasoning processes needed for deeper learning and long-term success.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_ea63748569", "k": "Expert Behavior Prior Reinforcement Learning", "domain": "science", "knowledge": "Behavior prior reinforcement learning (BPRL) has emerged as a promising paradigm to improve sample efficiency in online reinforcement learning (RL) by leveraging policy priors derived from offline demonstrations. However, most existing BPRL methods rely on static offline datasets, which often suffer from low data diversity and suboptimal trajectory quality. This reliance restricts the effectiveness of policy priors, hindering both policy exploitation and stability during online training. Consequently, agents are prone to inefficient exploration and unstable learning dynamics. To address these limitations, we deviate from existing offline pre-training methods and propose an Expert Behavior Prior (EBP) algorithm. Specifically, we introduce a Q-guided conditional variational autoencoder (Q-CVAE) that learns to generate expert policy priors directly from the online replay buffer. This enables the generation of high-value actions for guiding policy updates without relying on pre-collected expert trajectories. To further enhance policy exploitation, we propose an expert policy guidance (EPG) mechanism that selects expert actions from a generative support set, and we integrate a policy gradient correction (PGC) module to harmonize Q-guidance with expert supervision, promoting stable and consistent policy improvement. Extensive experiments conducted on robotic control (Gym, PyBullet) and industrial control (DMControl) benchmarks demonstrate that EBP significantly outperforms state-of-the-art online RL algorithms, achieving higher sample efficiency and more stable convergence.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_b9365cc53f", "k": "Unlearning Under Imbalance: Benchmarking Fairness in Multimodal LLM Unlearning", "domain": "science", "knowledge": "Machine unlearning has emerged as a tool for removing personal data from trained models to comply with recent AI regulations. To evaluate unlearning effectiveness in multimodal large language models (MLLMs), prior works fine-tune models on fictitious identities, simulating unlearning requests on subsets of these IDs, which are typically uniformly distributed. However, in realistic scenarios, people from different demographic groups may request to be unlearned at different frequencies, potentially altering the model's internal beliefs for these groups and leading to biased behaviors. To fill this gap, we propose FAIRGET, the first Visual Question Answering benchmark that evaluates unlearning under unbalanced, realistic, forget requests. These requests are designed to simulate multiple realistic scenarios, ranging from simple to challenging settings, that lead to biased unlearned models if fairness is not accounted for. Additionally, we propose FAUN, the first unlearning algorithm for MLLMs that forgets unlearning data while preserving model fairness. FAUN exploits a bias-aware activation steering mechanism to unlearn identities while accounting for the unbalanced nature of the forget data. Experiments on FAIRGET and the established FIUBench demonstrate our method's superiority both in unlearning quality and fairness.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_5d7b454e4b", "k": "An LLM-Driven Workflow for Automated Process Control Strategy Generation and Tuning from Dynamic Process Models", "domain": "science", "knowledge": "We present a structured large-language-model-driven workflow for automated multi-variable control design from dynamic process models. The workflow decomposes the design task into constrained code-generation steps: plant-interface construction, normalization, manipulated-variable controlled-variable (MV-CV) pairing, controller specification, closed loop simulation, scenario generation, performance evaluation and Bayesian-optimization (BO) based tuning. Generated artifacts are executed and validated before downstream tasks proceed, and failed artifacts are repaired using validation feedback. The proposed approach is demonstrated on a nonlinear gas-preheater benchmark with coupled pressure and temperature dynamics. The generated workflow produces a physically consistent decentralized PI (proportional-integral) feedback-feedforward control structure and an executable tuning environment. Bayesian optimization reduces the closed loop performance objective, which aggregates set-point tracking and disturbance-rejection errors for the controlled variables, by approximately 26.5% relative to the initial controller generated by the workflow, mainly through improved pressure-loop transient performance. This figure quantifies the automated tuning stage rather than a comparison against a manually designed controller. The results demonstrate the feasibility of using structured large-language-model-based code generation to construct executable control-design workflows, while also highlighting the need for broader validation on larger plantwide-control benchmarks.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_6ea09f814e", "k": "Multi-Task Learning for Heterogeneous Prediction from Video Game State with Transfer Learning", "domain": "science", "knowledge": "Multi-task learning (MTL) is a promising approach for prediction tasks derived from video game state data, as modern game telemetry provides multiple related supervision signals from the same structured observations. We study whether a shared model trained jointly across tasks in team-based multiplayer games can improve generalization while reducing training and inference cost compared to specialized single-task models. We adapt a multimodal architecture for endpoint prediction to a general multi-task setting that combines rasterized vision inputs, global match context, and per-unit state information through an image encoder and attention-based interaction modeling. Experiments on a large proprietary World of Tanks dataset compare single-task and multi-task training, evaluate weighting strategies for mixed losses and conflicting gradients, and test pre-training/fine-tuning under limited target-data regimes. We also examine within-game transfer across game maps under structured environment shift.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_e9abd99b8e", "k": "A Comparative Evaluation of Embeddings and LLMs in a Greek Book Publisher Setting - The CUP Dataset", "domain": "science", "knowledge": "We present CUP, a Greek book retrieval benchmark consisting of 868 catalog records and 104 expert-annotated queries with graded relevance judgments. We evaluate sparse (BM25), dense (sentence-transformers), hybrid, and LLM-assisted retrieval methods in this book-search setting. Multilingual embeddings outperform Greek-specific models, while hybrid retrieval performs best overall. A query-level analysis shows that BM25 excels at named-entity queries, while dense and hybrid methods improve natural-language, noisy, cross-lingual, and concept queries. Field-aware prompting has model-specific effects, while LLM TOC summarization improves TOC-only retrieval and LLM post-filtering improves early-stage retrieval at a high cost. Overall, CUP enables real-world evaluation of Greek retrieval across lexical, semantic, noisy, and cross-lingual queries.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_58433d3379", "k": "pAI-Econ-claude: A Gated Human-in-the-Loop Multi-Agent Architecture for AI-Assisted Economic Theory Development", "domain": "science", "knowledge": "In many social-science research tasks, such as economics, LLM-based agents must produce outputs for which no cheap, task-complete, machine-readable correctness signal exists. This creates a distinctive reliability problem for multi-agent systems: how should generation, critique, coordination, and human judgment be organized when no component can certify the final result? We address this problem through pAI-Econ-claude, a gated, human-in-the-loop multi-agent architecture for AI-assisted economic theory development. Agents coordinate through a shared workspace of inspectable intermediate records; specialized gates diagnose targeted failure modes and recommend loopbacks without certifying correctness; and human checkpoints retain authority over decisions that are costly to reverse. We evaluate the architecture on five matched economic-theory tasks against an ungated baseline. Two evaluators blinded to configuration agreed on all five pairwise rankings, preferring the gated architecture in four tasks and the baseline in one. Mean failure severity fell from 1.58 to 1.16, while overall usefulness rose from 2.60 to 3.10. The largest observed gain occurred when a reality check rejected a false market-structure premise and a proof review prompted revision of a false welfare claim. The negative case shows that scaffolding can also compress an economically important mechanism too aggressively. The results support a bounded claim: gated oversight improves the auditability of AI-assisted economic theory without substituting for formal verification, and the allocation of irreversible human judgment is a more informative design variable than pure agent autonomy. The workflow is publicly available at https://github.com/maxwell2732/pAI-Econ-claude.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_66f5d91480", "k": "BasketEvent: Understanding Who Did What and When in Basketball Videos", "domain": "science", "knowledge": "Comprehensive basketball video understanding requires resolving not only what event occurs, but also who is responsible and when the key evidence appears. However, exist- ing methods typically treat spatial perception and semantic recognition as isolated tasks, failing to ground events to individual players or pinpoint their temporal boundaries within complex collective dynamics. To bridge this gap, we introduce BasketEvent, a player- centric basketball event understanding dataset curated from real NBA broadcasts. In BasketEvent, event labels are grounded to the responsible players, and a manually an- notated subset of 1,000 samples with precise event intervals is provided to evaluate tem- poral evidence localization. Based on this data, we propose PlayNet, a player-centric reasoning framework that maps basketball videos to player-level event predictions with temporal evidence. Concretely, PlayNet tracks key entities, associates player identities, and reasons about events by modeling player-player, player-ball, and global court inter- actions, while aggregating sparse temporal evidence via gated pooling. Extensive experi- ments demonstrate that PlayNet significantly outperforms representative video-level and crop-based baselines, proving the superiority of player-centric modeling for fine-grained sports video understanding. Our data, code, and models will be made publicly available.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_8190b46601", "k": "slang.gr as a Large-Scale Crowdsourced Resource for Non-Standard Greek", "domain": "science", "knowledge": "Slang is a central component of everyday language, reflecting linguistic creativity, social identity, and cultural change, yet its dy- namic and non-standard nature makes it difficult to model computationally. We present the first large-scale computational study of slang.gr, a crowdsourced lexicon of Greek non-standard language, combining lexical content, user-generated tags, and interaction data. To enable the systematic analysis, we map noisy folksonomic tags to a structured multi-layer taxonomy capturing both semantic categories and sociolinguistic metadata. Using this representation, we analyze the linguistic structure of Greek slang and the behavior of its contributor community. We find that slang is strongly centered on person-related and evaluative language, exhibits high morphological creativity, and is shaped by highly skewed participation with short user lifespans and overlapping communities. Building on these signals, we introduce a community-based confidence score for definitions that integrates user roles, interaction patterns, and moderation signals. Our results show that taxonomy-based representations improve interpretability while retaining meaningful aspects of behavioral structure, enabling a more structured and interpretable analysis of confidence signals. Overall, this work establishes slang.gr as a computational resource for non-standard Greek and provides a foundation for sociolinguistic NLP, bias analysis, and the study of informal language in LLMs.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_ddb42571e0", "k": "Logic Programming Semantics for Causal Processes", "domain": "science", "knowledge": "Motivated by challenging modelling issues in the life sciences, we investigate the relationship between logic programming semantics and the eventual states of causal processes compatible with those logic programs. More precisely, we show that while stable models of positive logic programs correspond to the eventual states of processes commencing from a neutral state and continuing undisturbed indefinitely, supported models describe the eventual states reachable from arbitrary starting points. This also contributes to the discussion of the appropriate semantics for logic programming as a causal rule language, adding a temporal perspective to recent interpretations of the stable and supported model semantics from an explanatory viewpoint of causality.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_5dca845878", "k": "ICAE-Bench: Evaluating Coding Agents as Interactive Project Builders", "domain": "science", "knowledge": "The recent emergence of vibe-coding workflows is changing what coding agents are expected to do. Instead of merely completing code under fully specified instructions, agents are increasingly expected to transform incomplete product intent into working software by combining various abilities including planning, requirement clarification, tool use, debugging, and repository-level construction. Yet existing benchmarks have not fully caught up with this shift, evaluating agents on static, fully specified tasks. In this paper, we introduce ICAE-Bench, a benchmark for evaluating coding agents under interactive project-building settings. The basic idea is to start from a fuzzy product requirement, simulating the dynamic paradigm with an automated User Agent. To make this setting both realistic and evaluable, ICAE-Bench introduces three key designs. First, to avoid the ambiguity of unconstrained fuzzy requirements, each task derives ambiguity from a precise real open-source repository with executable behavior. Second, to ensure high-quality and reproducible user simulation, ICAE-Bench grounds interaction through User Agent Data, allowing the User Agent to reveal hidden constraints without inventing new requirements or leaking implementation artifacts. Third, to evaluate open-ended repositories fairly, ICAE-Bench uses standardized black-box tests together with multi-dimensional diagnostics, including functional correctness, semantic and API similarity, structural fidelity, design quality, and interaction quality.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_0df8147e6c", "k": "Explainable Belief Harmonization under Dynamic Epistemic Partitions", "domain": "science", "knowledge": "Existing approaches to multi-agent belief combination have established mature foundations for combining uncertain beliefs under common assumptions: consensus methods use iterative averaging, logic-based methods resolve conflicting knowledge bases, and epistemic logic analyzes agents' information states. Typically, these approaches assume that the structure determining what each agent can represent remains fixed. However, in many scenarios, agents gain or lose observational capacity during execution, and what was once admissible may become structurally impossible. This paper presents a formal framework for handling such runtime changes in epistemic partitions over continuous belief profiles. A hybrid approach exploits the advantages of answer set programming in elaboration tolerance, declarative integrity constraints, and explanations, with the numerical flexibility of Python. The framework applies to domains where agents operate at heterogeneous and possibly changing levels of resolution, and provides formal guarantees of admissibility preservation under refinement, unique mass-preserving repair under coarsening, and explanation completeness. Evaluation across 100 randomly generated topology changes confirms complete violation detection and explanation coverage.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_373887e844", "k": "Explainability Framework for Policy-Aware Autonomous Agents", "domain": "science", "knowledge": "In the field of Artificial Intelligence, an agent is a system which is able to autonomously make decisions in order to reach a desired goal. As these systems grow more prevalent in our day-to-day lives, there has been an increased need to add explainability features which can provide an account for an agent's behavior. We therefore propose a framework that outlines how to produce comprehensible explanations for policy-aware agents, or agents which have rule-enforcing policies incorporated in their decision-making framework. This framework is designed using insights from the social sciences on how to produce good explanations. It is implemented in the Answer Set Programming language while using Python to assist with information extraction and natural-language translation. Because these agents incur penalties when violating policies, we are able to leverage these penalties to detect undesirable events in scenarios that are counterfactual to the agents' original actions. This lends itself to creating contrastive explanations (e.g., \"the agent performed this action because, had it not, undesirable event X would have occurred.\"), which formulate the core component for our explainability framework. The framework is evaluated using a survey wherein human participants provide feedback on our program-generated explanations.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_2717da0686", "k": "How Rules Represent Causal Knowledge: Causal Modeling with Probabilistic Logic Programming", "domain": "science", "knowledge": "Pearl famously argues that causal knowledge enables the prediction of intervention effects. By contrast, purely descriptive knowledge supports only conclusions drawn from observations. His theory of causality, however, is developed exclusively within Bayesian networks and causal models. Consequently, it is largely restricted to acyclic causal relationships, and transferring its ideas to other formalisms risks misinterpretation or inconsistency. This paper brings Pearl's approach to causality into probabilistic logic programming (PLP). To this end, such programs are aligned with philosophical foundations established in prior work that do not rely on temporal notions; that is, all relevant events are assumed to occur simultaneously. A formal causal semantics for these programs, together with a notion of intervention and an implementation, is proposed. It is shown that this semantics coincides with the P-log semantics for stratified ProbLog programs, while the two may differ in the non-stratified case and for other PLP formalisms.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_17220c0595", "k": "A New Well-Supported Semantics for Description Logic Programs", "domain": "science", "knowledge": "Description logic programs are a powerful formalism for combining rules with ontologies. The well-supported semantics for description logic programs ensures that no answer sets rely on cyclic dependencies. Most popular semantics for logic programming have this property of well-supportedness. We recognize two limitations of the current well-supported semantics for DL programs: its increased computational complexity for the consistency problem and its lack of a reduct transformation characterization. In this work, we present a new semantics which evaluates ontological atoms more strictly than the current semantics. This keeps the complexity of its consistency problem NP-complete, rather than increasing it to the second level of the polynomial hierarchy. Additionally, we identify a syntactic class of description logic programs for which our new semantics is equivalent to the current semantics. We characterize our semantics using a fixpoint operator and a reduct-based transformation. Our new semantics is a strict subset of the current well-supported semantics, so it maintains the prior notion of well-supportedness while inducing its own stricter notion. We prefer our new notion of well-supportedness due to its similarities with logic programming.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_a8d15aff46", "k": "Hybrid MKNF with Classical Negation in the Rule Component", "domain": "science", "knowledge": "Hybrid MKNF knowledge bases under the well-founded semantics integrate Description Logics with Logic Programming. However, they do not support classical negation in the rule component, limiting their ability to represent explicit negative knowledge. This limitation is particularly significant in safety-critical applications, where reasoning often requires explicit negative information rather than interpreting the absence of information as evidence of absence. To address this issue, we introduce an extension of Hybrid MKNF that supports classical negation in the rule component. We formally define the syntax and semantics of the extended language and present a general procedure for computing its well-founded model.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_b94873c34d", "k": "Bound-Founded Semantics for Answer Set Programming with Difference Constraints: Preliminary Report", "domain": "science", "knowledge": "While the integration of linear constraints has significantly expanded the reach of Answer Set Programming (ASP), existing hybrid solvers often rely on disparate semantic underpinnings that lack a unified logical foundation. We address this gap by introducing a many-sorted variant of the Bound-founded Logic of Here-and-There (HTb), providing a versatile framework capable of characterizing equilibrium models across a wide spectrum of alternative semantics for extensions of ASP with linear constraints. We apply this framework to the setting of difference constraints, focusing on the semantic characterization of clingo[DL]. Central to our approach is the formalization of foundedness for numeric variables. By investigating how different hybrid systems - such as clingo[DL], clingcon, and flingo - justify constraint atoms, we uncover the semantic roots of their varying behaviors. This investigation results in a single, consistent framework that not only formalizes the foundations of current systems like clingo[DL] but also facilitates the rigorous study of program simplifications and the future integration of diverse semantic principles.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_90ed6131da", "k": "Towards a Certifying Grounder", "domain": "science", "knowledge": "Grounding, the translation of high-level theories into equivalent quantifier-free formulas, is a crucial step in declarative solving, yet it has so far escaped the proof-logging revolution. When this grounding step is not certifying, there is no way of knowing that the obtained solutions actually correspond to the original problem specification, resulting in a trust gap. In this paper, we close the trust gap between the user's high-level specification and the solver's low-level input by introducing a novel certifying grounding framework for first-order logic model expansion (FOX) over finite domains. We present CertiFOX, a framework consisting of: (1) a proof format for grounding derivations, (2) GroundFOX, a certifying grounder operating on theories in Grounding Normal Form (GNF)--a new normal form designed for compact, domain-aware grounding--and (3) CheckFOX, an independent proof checker. Our approach guarantees that the grounder's output is equivalent to the input specification, setting the stage for trustworthy end-to-end certified solving pipelines for declarative languages. Experimental evaluation confirms that CertiFOX is a feasible approach. The GroundFOX grounder is broadly comparable with other grounders, and proof checking with CheckFOX adds overhead within a small constant factor of grounding time.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_7e0936cae7", "k": "Declarative Problem Solving in UAM Strategic Deconfliction", "domain": "science", "knowledge": "The growing demand for Urban Air Mobility (UAM) introduces significant challenges in airspace management, particularly within densely populated metropolitan regions. As the number of aerial vehicles-such as drones, air taxis, and helicopters-continues to rise, so does the risk of mid-air collisions and conflicts with existing air traffic and obstacles. Ensuring safe and efficient UAM operations requires robust strategic deconfliction mechanisms. We propose an Answer Set Programming (ASP) based approach for strategic deconfliction, focusing on time synchronization and route optimization for conflict-free flight plans. The solution is benchmarked against Constraint Programming (CP), emphasizing scalability and resource use. Results show that ASP offers faster execution and better scalability for small to medium cases, while CP maintains stable memory but degrades with complexity.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_cfecd11b15", "k": "Case study: solving P-99 with LPTP and an LLM", "domain": "science", "knowledge": "Ninety-Nine Prolog Problems (P-99) is a famous set of Prolog exercises. We solved the first thirty three just by prompting an LLM (Large Language Model). We used Claude from Anthropic. By solved we mean: generate the Prolog code and a test file, run the tests and check whether they pass, then formally prove types, groundness, termination, uniqueness, existence and also sometimes functional correctness with LPTP (Logic Program Theorem Prover). Hence our approach is an experiment in vibe-coding/vericoding of P-99. It is a vibe-coding experiment because we started from informal specifications written in English and let Claude generate the Prolog code. It also fits within vericoding because the LLM proved reliability guarantees on the generated Prolog code. Claude wrote 58 logic procedures, 508 tests, 257 lemmas for a total of 11800 proof lines. We manually checked each file generated by the LLM. We checked the Prolog code, ran the tests, examined the logical statements generated by Claude and proof-checked Claude's proofs with LPTP. This paper describes this experiment and provides the main details so that it can be reproduced by the interested reader.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_a8ad2677ae", "k": "Chess\\_db: A framework for working with large chess game datasets", "domain": "science", "knowledge": "Chess is a two player strategic game that is embedded in classical AI culture as it was once the frontier for intelligent behaviour. There was the silent assumption that the advent of computer engines that play better than the best humans will extinguish interest in the game. However, the opposite has come to pass, with a growing following for the game. A lot of the computational resources are now centered around training of players, where the engine output is just one aspect. Access to past games is also an essential part, both in knowing what games a specific player has played previously, and also which continuations at a certain position have led to victory more often for each of the two colour players. We present Chess_db a suite of logic programming tools that can effectively manipulate games both in memory and via creating back end databases. In particular, we provide versatile code that creates databases from PGN (portable game notation) game files and explore the suitability of open source key-value databases for storing position tables that provide near-instant access to information pertaining to substantially large number of games.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_9fe7bb120f", "k": "Animation, Verification and Visualisation of Prolog Transition Systems with ProB", "domain": "science", "knowledge": "ProB is a Prolog-based model checker, animator and constraint solver for high-level formal specifications. One can also use ProB to animate transition systems defined by Prolog predicates, allowing the application of its various validation techniques. In this work, we present the existing features of ProB's Prolog animation mode and its recent extensions. The extended capabilities include simulation for statistical checks, more reliable trace replay, transitions with user input and improved state visualisation. We apply the new features to case studies, particularly for evaluating different strategies in game play, such as Connect Four. The features are useful for many other applications, especially for ProB's new sequent prover for Event-B proof obligations, as well as for demonstration models for teaching in combination with interactive visualisation.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_b40c52a958", "k": "Encoding Event-B Proof Rules in Prolog: An Interactive Sequent Prover for ProB", "domain": "science", "knowledge": "Event-B is a formal method rooted in predicate logic and set theory. We encoded over 600 proof rules in Prolog, enabling a systematic, comprehensible proof analysis and construction. By integrating the proof rules into the Prolog-based validation tool ProB, we obtain an interactive proof system with proof tree visualisation. This has advantages in teaching, giving students direct control over the selection of proof rules. Our tool can import proof obligations from the Rodin platform and provides multiple exports: a trace file for proof replay in ProB, an interactive HTML document for tool-independent exploration of the proof tree, and an export back to Rodin, allowing the ProB prover to be used as second chain. Compared to the previous implementation of the proof rules in Java, the encoding in Prolog is more compact, maintainable and extensible. While a preliminary iterative deepening prover with simple heuristics is already available and useful for finding short proofs, we aim to obtain fast automatic provers in the future.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_ed11497a3a", "k": "Identifying Good Rules for Efficient SAT Encodings of Single-Constant Multiplication Using Machine Learning", "domain": "science", "knowledge": "The Single Constant Multiplication problem is a fundamental NP-hard optimization task in hardware design, which seeks to decompose a fixed constant using only additions, subtractions, and bit-shifts. Although dynamic programming methods can produce near-optimal SAT encodings for SCM, their encoding cost remains high for large constants. We propose a neuro-symbolic framework that accelerates SCM SAT encoding by identifying good rules for guiding operator selection during decomposition. Our approach employs a graph neural network model to predict promising operator types from constant decompositions, and exploits the resulting confidence scores to prune no-good choices in the symbolic search. Experimental results on unseen 17-32 bit constants demonstrate one to two orders of magnitude reductions in encoding time, over 97% reduction in memory usage, and an order-of-magnitude decrease in branching, while preserving near-optimal encoding quality in terms of additions. These results show that learning-guided symbolic strategies can significantly improve the scalability and efficiency of SCM encoding. Our code and data are publicly available at: https://github.com/Chufeng-Jiang/SCM_MLDP", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_a06bfe3c05", "k": "Case study: proving sqrt(2) irrational with LPTP and an LLM", "domain": "science", "knowledge": "We present the interactions with an LLM (Large Language Model) aiming at proving that the square root of 2 is not a rational number in an LP (Logic Programming) context. We start from a few basic pure logic programming predicate definitions. We rely on the LPTP (Logic Program Theorem Prover) system for stating and proving properties about logic programs. As the proof language of LPTP is based on natural deduction, the proofs are human readable. In our case study, we sketch in LPTP the usual proof showing the irrationality of the square root of 2. Then we describe the interactions we had with the LLM. We end up with a complete formal proof, partially generated by an LLM and fully proof-checked by LPTP.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_4fa4bf9df9", "k": "Differentiable Logic Programming to Mitigate Reasoning Shortcuts in Neurosymbolic Systems", "domain": "science", "knowledge": "Neurosymbolic (NeSy) systems integrate neural networks with logical reasoning to achieve both generalization and interpretability, but recent work has shown they are susceptible to shortcut reasoning behaviors. We propose a novel method using matrix-based differentiable logic programming to mitigate reasoning shortcuts in two phenomena: constraint satisfaction shortcuts, where constraints are satisfied without achieving the intended task, and cognition shortcuts, where biased data leads to semantically incorrect concept mappings despite logically sound inference. Building on recent matrix-based logic programming semantics, we introduce design elements to mitigate shortcuts, including a unified encoding of rules and constraints in a single matrix. We also identify connections to fuzzy logic t-norms and empirically compare their gradient flow properties. Through carefully designed experiments on MNIST variants, we show that one-to-one grounding of neural outputs to logical atoms significantly reduces both shortcut types compared to previous methods that rely on soft probability distributions. We then confirm that architectural choices in coupling symbolic knowledge with neural learning play a critical role in shortcut mitigation.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_361bfb6537", "k": "Explaining Weather Bulletins via ILP", "domain": "science", "knowledge": "Inductive Logic Programming (ILP) originated within the Logic Programming community in the Nineties as a framework for combining symbolic learning with declarative knowledge representation. Nowadays, mature ILP frameworks exist and they are capable of learning complex, non-monotonic hypotheses, thus broadening both the modeling capabilities and the scope of real-world applications of ILP. This work is primarily based on the FastLAS2 framework and aims to generate simple, interpretable hypotheses to help clarify the weather bulletins issued by OSMER FVG, the Regional Meteorological Observatory of the Italian region of Friuli Venezia-Giulia. In this paper we present a pipeline that, starting from simulated meteorological raw data and from OSMERs' bulletins (used as ground truth), extracts data as ASP facts and generates ILP examples. From such examples an explanatory hypothesis is then inferred via FastLAS2. Such a hypothesis (translated into natural language) explains the weather forecast issued by human experts, and in particular the rationale behind experts' choices of specific symbols in the bulletin pictogram (the symbol-annotated meteorological map of the forecast). The proposed approach is general, not specific to any particular region and it can equally be applied to bulletins from other sources and to different regions.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_f1b2e12e4a", "k": "Representative Sets in Propositional Abduction", "domain": "science", "knowledge": "The propositional abduction problem is a well-known form of non-monotonic reasoning where we are asked to find an explanation of a given manifestation. Recently, there has been an influx of results asking more refined questions about the solution space rather than only individual solutions. For example, we might be interested in finding two solutions that are sufficiently far from each other (diverse solutions) in the solution space. In this paper we consider a related representation question where we ask if a given set of explanations S can represent any other explanation (that is, whether their symmetric difference is smaller than a given k). We first study this problem from a classical complexity perspective and obtain a complete classification. While only a handful of cases are tractable, the increase in complexity compared to classical abduction is often smaller than expected. We then study the parameterized complexity for several parameters and obtain new tractable and hard cases. Interestingly, a full parameterized complexity classification would require resolving the parameterized complexity of the covering radius problem from coding theory. To the best of our knowledge, no useful relationship between coding theory and non-monotonic reasoning has previously been established, but such connections seemingly become important when asking more complex questions about solution spaces.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_e3a467f754", "k": "Safeguards for Speech2Speech LLM-Assistants: A Case Study in Automotive Applications", "domain": "science", "knowledge": "Recent advances have introduced speech-to-speech (S2S) conversational assistants capable of producing natural-sounding interactions, including non-verbal cues like tonality and mood. In the automotive domain, this enables intuitive and humanlike in-car dialogue experiences. However, integrating these end-to-end assistants limits architectural options for programmable domain-specific safeguards. This paper discusses two implementation approaches for S2S guardrails: transcript-based and tool-based. Through an empirical evaluation, we demonstrate that both strategies are insufficient for industrial deployment in most cases due to prohibitive latency (delaying each answer by 0 to 1.4 seconds even for computationally cheap checks) and technical impediments (like potentially non-deterministic tool call behavior). Finally, we outline open challenges for S2S guardrails in the automotive context.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_91e9ae7d80", "k": "SafeStep: AI-powered Travel Assistance for Elderly People with Frailty or Dementia", "domain": "science", "knowledge": "More than a million people in the UK suffer from frailty or dementia, which severely compromise their ability to travel in urban environments. This paper presents SafeStep, an AI-driven travel system that assists elderly users with their journeys. At the core of SafeStep is a novel travel graph representation, which integrates route planning with predictive modelling. For each stage of a journey, the system (i) generates personalized failure scenarios using a combi-nation of LLMs and the Anticip8 behavioral prediction engine, (ii) proposes targeted interventions, and (iii) estimates the impact of interventions on out-come probabilities. This enables SafeStep to select interventions that maximize the likelihood of the person reaching their destination. SafeStep was evaluated through experiments on travel graph generation and a field study involving 26 real-world journeys. Results showed that combining Anticip8 for failure pre-diction with GPT-based models for intervention evaluation yields the most re-liable performance. User feedback indicated that SafeStep improves confidence and perceived safety during travel, although interface usability needs to be im-proved for the target demographic. In the future, we would like to improve and release SafeStep. The AI system that was developed for SafeStep could be ap-plied in other areas, such as mental health, career coaching and addiction treatment.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_9276c24060", "k": "CRAG-MM-Diagnostics: Enabling Stage-Wise Analysis of Knowledge-Intensive VQA", "domain": "science", "knowledge": "Knowledge-Intensive Visual Question Answering (KI-VQA) benchmarks evaluate Vision-Language Models (VLMs) as multimodal knowledge assistants by requiring external information beyond a provided image to answer questions. KI-VQA involves multiple sub-problems -referring expression understanding, visual grounding, object recognition, knowledge retrieval, and reasoning-yet existing benchmarks typically report only end-task accuracy, obscuring where failures arise. To analyze the full KI-VQA pipeline, we introduce CRAG-MM-Diagnostics, a diagnostic benchmark with stage-wise data annotations that isolate 1) language-based visual grounding, 2) object identification, and 3) knowledge retrieval and reasoning. We evaluate fully parametric and retrieval-augmented VLMs, providing fine-grained analyses using newly collected metadata, such as target ROIs, entity names, and visual complexity scores. Our results point to knowledge retrieval and reasoning as the primary bottleneck, but also highlight issues in the other parts of the KI-VQA pipeline, such as the fact that VLMs struggle with target object identification or that image retrievers struggle to integrate textual cues. These findings expose fundamental limitations in current KI-VQA systems and motivate stage-aware evaluation. We, lastly, leverage these findings to propose a grounded bimodal RAG pipeline that integrates a visual grounding module to crop targets before image retrieval, boosting GPT-5 and Qwen's respective accuracies by 13.3 and 8.5 percentage points.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_5402450084", "k": "V-DEAL: Diagnosing Video Safety De-Calibration as an Understanding-Refusal Coupling Failure", "domain": "science", "knowledge": "As Video Large Language Models are increasingly deployed in real-world applications, ensuring their safety alignment has become critical. Counterintuitively, we find that harmful videos paired with benign queries achieve higher attack success rates than the same videos paired with explicitly harmful queries. To understand the underlying mechanism of this vulnerability, we present V-DEAL, a three-level diagnostic framework that jointly analyzes this failure across model behaviour, understanding, and internal representations. By progressively ruling out perception failure and quantifying the model's internal refusal tendency, V-DEAL provides a new diagnostic perspective for analyzing the underlying mechanism of the observed vulnerability. We tested six Video LLMs on three public benchmarks and observed that models correctly recognize harmful video content with over 81\\% accuracy, yet the average attack success rate still reaches 48.33\\% under the condition pairing harmful videos with benign queries. Hidden-state analysis further shows that visual understanding activates a weaker refusal tendency than textual understanding. Furthermore, we introduce a prompt injection intervention method that reduces attack success rates by an average of 48.24 percentage points and achieves performance comparable to prior fine-tuning-based methods, providing an effective and practical means to address such safety risks in Video LLMs.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_0070896983", "k": "One More Turn, Less Regret: A Regret-Based Multi-Turn Benchmark for LLMs' Clarification Policies", "domain": "science", "knowledge": "Ambiguous user requests make clarification a sequential decision problem for conversational LLM assistants: they must decide whether to ask, what to ask, when to stop, and when to answer. We introduce RegretBench, a multi-turn benchmark that evaluates clarification as policy behavior rather than isolated question quality. RegretBench provides a hidden-intent formulation of ambiguity, supports free-form interaction grounded in semantic-state tracking, and introduces a regret-based objective that measures how much value a model loses relative to a reference clarification policy. Experiments on open-domain QA and product recommendation scenarios show that final success alone is insufficient, as models with similar accuracy can differ substantially in efficiency, robustness to user behaviors, and stopping decisions. By jointly measuring intent resolution, interaction cost, ineffective clarification, and regret, RegretBench reveals whether models clarify usefully and efficiently. Our results show that effective clarification requires more than plausible questions: models must ask the right question at the right time and stop once the user's intended meaning is clear.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_ebafb88220", "k": "Demographically-Informed Heat-Mortality Risk Curves via Risk Graph Neural Networks", "domain": "science", "knowledge": "Estimating heat-related mortality risk is a core task in environmental epidemiology, typically addressed with Distributed Lag Non-linear Models (DLNMs); interpretable exposure-response surfaces fitted to temperature-mortality time series. DLNMs are effective but ignore demographic and geographic context, despite well-established relevance to heat vulnerability. We propose Risk Graph Neural Networks (RGNNs), a hierarchical GNN encoder that uses granular census features to optimise DLNM coefficient vectors, preserving interpretable risk curve outputs while substantially improving predictive calibration. Evaluated across 10 regions of England and Wales on two unprecedented heat years, RGNN variants maintain both lower point-errors and near-nominal uncertainty coverage during the 2022 heatwave where baselines collapse.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_22209317f3", "k": "Hardware-Software Co-Design for Float16 On-Device Training on RISC-V Single-Core", "domain": "science", "knowledge": "By leveraging standard RISC-V extensions, namely Zfh (scalar float16) and Zvfh (vector float16), this work proposes an open-source framework to enable complete on-device training on resource-constrained RISC-V single-core. Our approach allows memory footprint reduction by about 50% as compared to using float32 and with minimal model performance degradation. We also facilitate transfer learning and fine-tuning scenarios by incorporating layer-freezing capabilities. Our work builds onto AIfES, an open-source, modular and generic DNN training and inference framework for embedded systems that can be extended with custom hardware-specific functions. The benefits of float16 is further emphasized by outlining the low area overhead of Zfh on a RV64GC super-scalar out-of-order FPGA softcore (+1.15% LUT6 and +0.05% FF at 175MHz). Finally, we discuss the architecture of a Zvfh implementation within the same RISC-V core.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_b969d39a1e", "k": "Relative Value Learning", "domain": "science", "knowledge": "In reinforcement learning, critics typically estimate absolute state values $V(s)$, estimating how good a particular situation is in isolation. However, it turns out that only differences in value are relevant for control. Motivated by this, we propose Relative Value Learning (RV), a framework that learns value differences directly via an antisymmetric function $Δ(s_i, s_j) = V(s_i) - V(s_j)$. We introduce a pairwise Bellman operator and prove it is a $γ$-contraction with a unique fixed point equal to the true value differences, derive well-posed $1$-step, $n$-step and $λ$-return targets and reconstruct generalized advantage estimation from pairwise differences to obtain an unbiased policy-gradient estimator (R-GAE). Beyond theoretical results, we integrate RV with PPO and achieve competitive performance on the Atari benchmark (49 ALE games) compared to standard PPO, indicating that relative value estimation is an effective alternative to absolute critics.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_474b79d22a", "k": "GlucoTune: A Unified Framework for Blood Glucose Preprocessing, Forecasting, and Benchmarking in Diabetes", "domain": "science", "knowledge": "Preprocessing blood glucose time-series data is a critical yet often overlooked step in developing data-driven methods for diabetes management, particularly for type 1 diabetes. The lack of standardized preprocessing workflows and evaluation protocols hinders reproducibility and complicates fair comparison across studies. These challenges are further exacerbated by data-sharing restrictions, as privacy and licensing constraints often prevent the redistribution of preprocessed medical datasets. To address these limitations, we present GlucoTune, a comprehensive and extensible framework for reproducible experimentation with blood glucose time-series data. The framework standardizes the entire experimental workflow, from preprocessing to model evaluation, enabling reproducible experiments directly from the original datasets. Reproducible preprocessing is achieved through configurable pipelines defined in portable YAML configuration files, ensuring consistent data handling without distributing sensitive preprocessed data. Beyond preprocessing, GlucoTune provides a unified interface for implementing, training, and evaluating blood glucose prediction models. The framework integrates public datasets through standardized wrappers and provides a curated collection of state-of-the-art blood glucose prediction and general time-series forecasting methods, while remaining readily extensible to additional datasets, preprocessing strategies, and forecasting models. To promote transparent and consistent evaluation, GlucoTune includes a benchmarking leaderboard that reports results across datasets, preprocessing configurations, and forecasting methods, enabling systematic comparison of experimental settings. We demonstrate the effectiveness of GlucoTune through comprehensive experiments and assess its usability in a user study.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_0109141cdf", "k": "TOUR: A Trajectory-Level Unlearning Benchmark for Offline Reinforcement Learning", "domain": "science", "knowledge": "Offline Reinforcement Learning (RL) agents are trained on fixed behavioral trajectories, which makes trajectory-level deletion important when selected data must be removed after training. Evaluating such deletion is difficult because a lower membership score can reflect trajectory removal, residual memorization visible to another attack, or policy collapse that destroys useful behavior. We introduce Trajectory-level memOrization and Unlearning in offline RL (TOUR), a benchmark that combines trajectory-level partitioning, matched non-member controls, retraining references, retained-performance anchors, and multi-attack privacy auditing. Across D4RL locomotion experiments and an exploratory AntMaze extension, TOUR shows that common deletion baselines have environment-dependent privacy-utility behavior. Retraining and fine-tuning often provide stronger retained-utility references than uniform GA+Refit, while TrajDeleter remains a useful comparator but is not uniformly stronger under the same audit. Reference-model, threshold, deviation, equivalence, action-error, representation-based, and query-limited attacks further show that a single likelihood-based membership score can overstate deletion quality. In the evaluated settings, conclusions about offline RL unlearning are therefore not stable under single-score auditing. They depend on matched non-member construction, retraining-relative calibration, attack family, retained utility, and explicit scope for diagnostic architecture or component-level evidence.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_6f13fabc8c", "k": "AttriMem: Attribution-Guided Process Feedback for Agent Memory Learning", "domain": "science", "knowledge": "Effective memory is crucial for LLM agents, yet constructing it effectively remains challenging. A memory-construction policy decides what information to extract, store, update, compress, or discard as interactions accumulate. Heuristic memory methods rely on subjective, task-specific rules, which can misalign with downstream objectives and limit cross-task adaptability. RL-based methods, by contrast, learn from task feedback but mainly use outcome- or module-level rewards. These coarse signals indicate task success but cannot identify which intermediate memory contents support the final answer, creating a fine-grained credit-assignment bottleneck. However, constructing such process feedback is prohibitively difficult because intermediate memory decisions lack unique ground-truth targets, while the appropriate credit varies with the agent's uncertain reasoning trajectory and therefore cannot be specified in advance. We propose AttriMem, an attribution-guided process-feedback framework for learning memory-construction policies with RL. AttriMem augments the global outcome reward with local rewards derived from token-level contributions to the final answer. Experiments on long-horizon dialogue question answering show that AttriMem outperforms retrieval-based, heuristic, and RL-based baselines, generalizes across benchmarks and answer models, stabilizes RL optimization.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_4d363538d5", "k": "Can Generative Recommendation Reach Cold Items? A Temporal Perspective on Semantic-ID Generation", "domain": "science", "knowledge": "Semantic-ID-based generative recommendation represents items as sequences of shared semantic tokens, enabling token recombination beyond isolated item IDs. However, closed-world recombination does not necessarily imply temporal open-token cold-start induction, where new items enter the item catalog with unseen atomic tokens or weakly supported SID paths. In this work, we revisit SID-based generative recommendation under an absolute-time temporal protocol that separates seen and unseen targets and diagnoses the cold item reachability at the token level. Through seen/unseen-hit analysis, coldness taxonomy, and oracle-prefix probing, we show that current SID-based models can occasionally reach future items supported by observed tokens and prefixes, but struggle with unseen atomic tokens and unsupported SID paths. We further explain this boundary by interpreting SID generation as hierarchical semantic bucketing: early tokens select coarse semantic regions, while later tokens refine item-specific paths. These findings show that SID generation is compositional but not fully open-ended, and suggest future directions in more independent SID spaces, scoring-based interfaces, and dynamic textual context.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_0bcef3705d", "k": "Training Large Language Models for Self-Explanation Faithfulness", "domain": "science", "knowledge": "We propose a Reinforcement Learning (RL) method to directly optimize the faithfulness of self-explanations - the extent to which a model's generated reasoning accurately reflects its internal decision-making process. While existing work focuses on evaluating faithfulness or using inference-time prompting frameworks to improve an LLM's self-explanation's tractability, these approaches do not provide a mechanism to directly optimize a model's parameters to generate faithful self-explanations. We bridge this gap by modifying existing faithfulness metrics into an RL training objective. We investigate (1) if models can be trained to accurately detect factors that affect their decisions, and (2) whether RL can directly optimize for the disclosure of these factors thereby improving LLM self-explanations' faithfulness. We experiment with two intervention types: random-word insertions and user-bias insertions, using a per-sample reward derived from the Phi-CCT correlation metric. RL fine-tuned Llama3.1-8B and Qwen3-8B show substantial improvements on the Phi-CCT faithfulness metric, with in-distribution scores rising from near-zero to as high as 0.664, and out-of-distribution scores reaching up to 0.691 on held-out tasks such as StrategyQA. Cross-intervention generalization is weaker but more interesting: a priori we would not expect a model trained only on random word insertions to generalize to user-bias phrases, yet Llama3.1-8B shows non-zero transfer in this direction. The reverse direction and Qwen3-8B do not replicate this, indicating model-dependent and setup-dependent effects we cannot yet explain. Lastly we analyze model behavior to rule out reward gaming behaviors that often plague RL training. Ultimately, we show that models can be trained to implicitly identify influential factors and disclose them, offering a scalable path toward reducing unfaithful reasoning in LLMs.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_06bbd1275d", "k": "Faster IndexTTS-2: Accelerating and Streaming Autoregressive Zero-Shot Text-to-Speech Synthesis on GPUs", "domain": "science", "knowledge": "Autoregressive text-to-speech models achieve strong naturalness but suffer from slow inference due to sequential token generation, limiting their deployment in production applications that require low latency. IndexTTS-2 is a state-of-the-art autoregressive TTS model consisting of a GPT, a flow-matching Diffusion Transformer, and a vocoder. Despite its high synthesis quality, its inference speed barely reaches real-time without streaming or batching support. We present Faster IndexTTS-2, which accelerates all neural network components of IndexTTS-2 for production deployment on GPUs using NVIDIA TensorRT and TensorRT-LLM. Faster IndexTTS-2 also enables streaming synthesis for latency-sensitive interactive applications, and batched inference across all components to maximize GPU utilization. Experiments on the Seed-TTS benchmark for both English and Chinese demonstrate up to 5.0$\\times$ speedup on the autoregressive GPT and 3.6$\\times$ end-to-end, with minimal degradation in word error rate, speaker similarity, and naturalness. Our methodology provides a practical reference for efficiently accelerating similar autoregressive speech models on GPUs.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_a59044c24d", "k": "HiMe: Real-Time Self-Hosted Personal Agent Platform for Health Insights with Wearable Devices", "domain": "science", "knowledge": "Traditional approaches to wearable health signal analysis, such as smartwatches, are constrained by rigid analytical frameworks and limited personalisation. The emergence of LLM agents creates a new opportunity for Personal Health Agentic Analysis, where health insights can be generated adaptively and in context. However, currently there is no open-source locally deployable platform capable of processing personal health data in real time while preserving privacy. We present HiMe, a locally deployable, privacy-first agent platform that is fully compatible with real-time health data ecosystems across a wide range of wearable devices. HiMe is guided by three design principles. The database is treated as a first-class component. Effectiveness and efficiency are jointly optimised to achieve a low-cost Pareto-optimal balance. Data are processed in real time while the user is modelled over the long term. Together, these principles make it practical for individuals to harness Personal Health Agents for continuous, personalised health monitoring for better wellbeing.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_a11691f5ac", "k": "EmoAgent-R1: Towards Multimodal Emotion Understanding with Reinforcement Learning-based Dynamic Agent Specialization", "domain": "science", "knowledge": "Multimodal large language models (MLLMs) have achieved impressive performance in multimodal emotion recognition (MER) tasks and lifted MER to a new level that is complex emotion understanding with advanced video understanding abilities and natural language description. However, existing MLLM-based methods often use a fixed prompt to perceive the emotions, ignoring the dynamicity and complexity of the emotion source in the multimodal inputs. To address these issues, we propose a novel Reinforcement Learning-based Dynamic Agent Specialization framework (\\textbf{EmoAgent-R1}) to optimize the emotion recognition, reasoning, and generalization abilities of an MLLM with dynamic agent specialization based on reinforcement learning. Specifically, we first adopt a cold start strategy to endow an MLLM with preliminary emotion recognition, reasoning, and agent routing ability by training with synthetic answer-conditioned chain-of-thought data and agent routing data. Then, we further train the MLLM with reinforcement learning to perceive emotions in a two-step agentic workflow with agent selection and agent specialization. To effectively train EmoAgent-R1, we propose a novel Progressive Group-Relative Policy Optimization (P-GRPO) to combine group-based relative advantages with a PMI-inspired progressive token-level modulation to transform sparse rewards into fine-grained learning signals, mitigating the coarse-grained uniform credit assignment issue in GRPO. Extensive experiments on MER benchmarks demonstrate the superiority of our EmoAgent-R1 in stronger emotion reasoning performance and improved optimization stability.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_888e382ee6", "k": "Reexamining zero-shot summarization: Empirical investigation of trustworthiness of LLM-summarizers", "domain": "science", "knowledge": "Zero-shot summarization using Large Language Models (LLMs) has significantly advanced the abstractive summarization task by producing coherent and fluent summaries. However, underlying stochasticity of the large language models raises concerns about the stability and trustworthiness of the LLM-generated summaries. This issue has become increasingly important due to proliferation of LLM-generated summaries in educational settings, where students and researchers summarize complex academic materials in zero-shot manner. We propose a novel two-level diagnostic protocol for benchmarking LLM-summarizers based on the stability of the generated summaries. At the lower level, document-level stability analysis is performed over multiple LLM-summaries generated under controlled environment, and the stability coefficient is computed. Each generated summary is scored for semantic and factual alignment with the original document, enabling estimation of stability along more than one dimensions. At the next level, observations from a stratified sample of documents drawn from the corpus are consolidated to estimate the stability index of the LLM-summarizer, which is the proxy for its trustworthiness. Our empirical investigation of three LLM-summarizers across three genres of documents reveals statistically significant differences in the generation-level variability among LLMs across summary evaluation metrics. This study advances the LLM-summarization research by evidential recognition of the stability problem in LLM-summaries and motivates further research towards development of robust, reliable and trustworthy LLM-summarizers.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_31a72254ab", "k": "Naju: A Native Discrete State-Space Model with Independent Retention and Writing for Long-Sequence Memory", "domain": "science", "knowledge": "Long-sequence memory tracking places two opposing demands on a recurrent state: near-lossless retention of stored bindings over long horizons, and active overwriting of stale ones. In our diagnostic suite, the strongest efficient baselines tend to solve only one side well. Continuous-time-parameterized state-space models (SSMs) such as Mamba obtain their discrete recurrence by zero-order-hold discretization of a continuous-time system; we argue that this detour is unnecessary for memory tracking and parameterize the discrete transition directly. Naju (Native Adaptive Junction Unit) factorizes the recurrent update, schematically $x_n = f_n\\odot x_{n-1} + i_n\\odot(B_n u_n)$, into an explicit discrete pole (a learned forget gate $f_n$), an independent write gain $i_n$, and input-dependent write/read maps. Since the sigmoid pole satisfies $0<f_n<1$, each frozen local coordinate is Schur-stable by construction, and the full time-varying recurrence satisfies a fading-memory/BIBO bound under uniform boundedness assumptions, with no stability regularizer. We formalize the key structural limitation of coupled designs: any non-expansive complementary single-gate recurrence ties the effective retention $r$ and write gain $w$ through $|r|+w\\le 1$, so near-complete retention forces weak writing; decoupling $f_n$ from $i_n$ removes this constraint. Empirically, Naju is the only evaluated model that remains strong on both retention and overwriting at 4x the training length. Beyond the diagnostic suite, we evaluate Naju on WikiText-103 language modeling, Long Range Arena, and multi-query associative recall. Across these settings, Naju consistently combines strong long-range memory with competitive or superior performance, outperforming the Mamba baselines in the principal comparisons while remaining competitive with the Transformer and preserving linear-time, linear-memory scaling.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_cbcb656442", "k": "Workflow-Localized Mechanism Learning: Attribution-Guided Repair and Knowledge Reuse for Structured Agent Skills", "domain": "science", "knowledge": "Agent Skills package reusable procedural knowledge as external artifacts for frozen language-model agents, yet existing optimizers do not jointly resolve where a failure occurs in a workflow, which mechanism caused it, and how relevant knowledge from third-party Skills should be reused locally. We introduce Workflow-Localized Mechanism Learning (WML). Its Node--Mechanism Attribution identifies the failed workflow node, implicated mechanisms, and smallest valid edit target, routing single-mechanism defects to L3 resources and relational defects across mechanisms to L2 composition protocols. A six-module Workflow-Guided Skill Optimization (WGSO) loop then selects provenance- and scope-aware third-party knowledge, applies bounded patches, evaluates candidates, and stores verified outcomes in optimizer-side memory. On SpreadsheetBench, WML reaches 90.33 +/- 1.53 and 74.67 +/- 3.51 Hard Accuracy with DeepSeek and Qwen3.6-Flash, respectively; without additional optimization, the learned Skills transfer to WikiTableQuestions with 84.00 +/- 2.00 and 83.00 +/- 2.00 Denotation Accuracy. On Compiler-Supported50, WML attains both the highest hard-PASS rate and the lowest cost per successful task; compiled execution sharply reduces tokens and calls relative to a direct SkillAgent while retaining most of its successful tasks. Code and artifacts are available at https://github.com/xiaolin9595/workflow-localized-mechanism-learning.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_396a62ea83", "k": "Sparse Concept Channels in Frozen 3D CT Vision Encoders", "domain": "science", "knowledge": "Large vision-language models are becoming increasingly dominant in 3D medical image interpretation, but we rarely know <i>which</i> internal units encode clinical findings or <i>where</i> that information lives in the representation. We first study this on a 3D chest vision-language model (Pillar-0) by probing its frozen vision embeddings. We show that (i) each radiological finding is encoded by a <i>sparse</i> set of ~10 vision-encoder channels that match full-feature classification performance and far exceed a zero-shot text prompting; (ii) turning off the channels tied to one finding, that finding's score collapses while unrelated labels stay stable; and (iii) the same sparse probe <i>replicates</i> on an architecturally unrelated 3D abdominal VLM (Merlin) suggesting a general property of frozen medical encoders. Our training-free concept channel probe (CCP) method, paired with a corpus-derived report template, outperforms published CT-CHAT on clinical efficacy and NLG metrics (F1 0.549 vs. 0.184; BLEU 0.483 vs. 0.373) at 22x lower latency. Our results provide a clear, reproducible characterization of how frozen medical encoders represent findings, demonstrating direct applicability across models.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_1bcd43e4a1", "k": "HyWorldVLA: A Vision-Language-Action Model with Hybrid World Modeling for Autonomous Driving", "domain": "science", "knowledge": "Vision-Language-Action (VLA) models augmented with world modeling represent a promising paradigm for end-to-end autonomous driving. While pixel-level future prediction enables fine-grained spatiotemporal reasoning, it compromises robustness in noisy driving scenarios. Conversely, latent-based world models alleviate this sensitivity but often incur limited interpretability and representational degradation due to absent pixel-level grounding. To reconcile this trade-off, we propose HyWorldVLA, a hybrid world-VLA framework that unifies pixel-level supervision and latent representation learning. In the pre-training stage, HyWorldVLA predicts video latents encoded by a pre-trained video VAE, while simultaneously reconstructing video frames to provide precise pixel-level grounding. During the subsequent co-fine-tuning phase, the model exclusively predicts latent features, which are fed into an action expert to generate trajectories. Extensive experiments on NAVSIM v1 and v2 benchmarks demonstrate that HyWorldVLA significantly outperforms both pixel-based and latent-based world model baselines. Notably, we present the first comprehensive qualitative and quantitative analysis of world model noise robustness in autonomous driving, establishing a new benchmark for evaluating future architectures.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_9904d42c25", "k": "3D-Aware VLMs with Implicit and Explicit Geometries", "domain": "science", "knowledge": "Despite rapid progress, most existing vision-language models (VLMs) built from 2D visual inputs often struggle when handling various 3D tasks that require fine-grained spatial understanding and reasoning. To bridge this gap, we present VLM-IE3D, a unified framework that enhances the 3D spatial awareness of VLMs by equipping them with both implicit and explicit 3D geometries learned from RGB videos. Our VLM-IE3D introduces Implicit Geometry Tokens (IGTs) that capture high-level geometric priors from input videos, as well as complementary Explicit Geometry Tokens (EGTs) that encode detailed geometric structures from reconstructed 3D attributes. On top of that, VLM-IE3D comes with a 3D-aware adapter that effectively fuses the two types of geometric representations with 2D visual cues. This RGB-only design injects strong 3D inductive biases for fine-grained spatial understanding and reasoning without requiring any additional 3D inputs. Extensive experiments show that VLM-IE3D achieves superior performance consistently across various 3D tasks including 3D video detection, 3D visual grounding, 3D dense captioning, and spatial reasoning. Code and models are available at https://github.com/Vegetebird/VLM-IE3D.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_d2f3fb7c86", "k": "Expanding Flow Maps", "domain": "science", "knowledge": "Flow-based generative models have enabled remarkable progress in fast and controllable generation across continuous and discrete state spaces, yet existing parameterizations are constrained to fixed dimensions or fixed sequence lengths. Here, we introduce Expanding Generative Flows (EFlows), which define flows between distributions of increasing dimensionality along an expanding interpolant that grows the state by augmenting it with conditional noise. Building on this construction, we propose Expanding Flow Maps (EFMs), a new class of flow maps that distill the expanding interpolant into efficient few-step generative models. Each EFM factors the map between any two timesteps into two learnable operations: an expand operator, which augments the state space with new coordinates or tokens conditioned on the current state, and a transport map, which pushes the expanded state forward along the interpolant. Composing these operators yields a single map that jointly expands and denoises the state, recovering existing fixed-canvas flows and flow maps as the special case in which the expand operator is the identity. We further extend the framework to the discrete simplex, enabling variable-size graph generation and variable-length sequence generation. Across both continuous and discrete modalities, we establish EFlows and EFMs as a principled framework for settings in which output size is itself a learned, controllable degree of freedom.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_a05a246486", "k": "Barzilai-Borwein Fails Superlinear Convergence on an Open Set of Quadratics for Every Dimension $n\\geq 4$", "domain": "science", "knowledge": "Barzilai--Borwein (BB) method has shown strong practical performance in continuous optimization, yet its convergence dynamics remains poorly understood. In particular, a central unresolved question is whether BB converges superlinearly for almost every strictly convex quadratic problem and initialization. We provide a negative answer to this question. Specifically, for every finite dimension $n\\geq4$, we construct a nonempty open, hence positive-Lebesgue-measure, family of strictly convex quadratic problems and initial points for which the long Barzilai--Borwein method (BB1) converges but cannot converge root-superlinearly. More precisely, with the explicit constants $ρ_{\\min}=10^{-6},ρ_{\\max}=0.61$, every spectral component of the gradient is bounded above and below by the corresponding geometric sequence. Consequently, the gradient norm and the energy norm of the error satisfy two-sided geometric estimates with the same rates, while the objective gap satisfies the corresponding estimates with squared rates. In particular, all three quantities are bounded below by geometric sequences, ruling out superlinear convergence. The construction is highly nontrivial, based on a computer-assisted proof of a nonresonant, attracting seven-cycle of the projectivized BB dynamics in dimension four.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_b82ad00c90", "k": "Synthetic data generation framework for quality control automation in gravure printing", "domain": "science", "knowledge": "Quality control in printing, particularly in rotogravure printing, still depends on slow, costly, and subjective manual inspection. Automated surface defect detection is critical for maintaining high-quality standards in rotogravure printing. Deep learning models give prospects for automation. However, training robust deep learning models, such as YOLO or Vision Transformers, is heavily hindered by the extreme scarcity of real-world industrial defects images. To overcome this limitation, this paper introduces a novel synthetic data generation framework tailored for rotogravure printing quality control. The proposed pipeline automatically generates high-fidelity images of specific printing defects (creases, streaks, misregistration, etc.) and outputs corresponding bounding boxes and annotations. To validate the framework, a synthetic dataset of 7533 images was generated and used to train the state-of-the-art object-detection model RFDETR. Experimental results demonstrate that the model trained on our synthetic data achieves a Mean Average Precision (mAP) of 80.9\\% on real industrial testing samples. This framework provides a zero-cost, rapid-deployment solution for automating defect inspection in printing lines without requiring massive manual data collection.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_928c2671ab", "k": "Beyond Sufficiency: Time Series Explanation with Counterfactual Necessity", "domain": "science", "knowledge": "Faithful explanations of time-series classifiers should identify subsequences that are not only sufficient to preserve a black-box model's prediction, but also necessary for maintaining it. However, existing sufficiency-oriented methods can assign high importance to spurious subsequences that support the prediction without being essential to the model's decision. We introduce \\textbf{TimePNS}, a necessity-aware framework for time-series explanation. Inspired by Pearl's counterfactual notion of necessity, TimePNS assesses whether a temporal factor is necessary by intervening on it and measuring whether the original prediction is disrupted. The framework adopts a two-stage design. Stage I learns an identifiable causal generative process together with a sufficiency-oriented explanation mask. Stage II performs counterfactual interventions on temporal factors to derive necessity signals, which supervise a temporal gate that refines the initial explanation by suppressing non-essential components and emphasizing counterfactually necessary ones. Experiments on synthetic and real-world time-series benchmarks show that TimePNS more accurately identifies decision-critical subsequences and consistently improves sufficiency-necessity trade-offs over strong baselines.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_854e4d8d2c", "k": "Graph Learning on Ensembles of Cyclic Peptides: An Investigation of Molecular Ensemble Modeling", "domain": "science", "knowledge": "Molecular property prediction from structure often uses a single representative conformation, even though many molecules exist as conformational ensembles in solution. We introduce EnsembleEGNN, a molecular ensemble foundation model that encodes an ensemble by first encoding each conformer with shared Equivariant Graph Neural Network (EGNN) layers, then pooling the resulting conformer representations with a Set Attention Block. We pretrain the model on CREMP, a cyclic peptide ensemble dataset, using a multi-task self-supervised objective combining masked token recovery, noisy-coordinate reconstruction, and pairwise distance reconstruction. On the CREMP-CycPeptMPDB dataset, training EnsembleEGNN from scratch fails entirely ($R^2=0.005$). However, the pretrained model reaches $R^2=0.477$ and Pearson $r=0.699$, outperforming the sequence-only BERT baseline ($R^2=0.439$, Pearson $r=0.667$). When EnsembleEGNN is co-trained end-to-end with the BERT sequence encoder, the hybrid model improves further to $R^2=0.538$ and Pearson $r=0.737$. These results demonstrate that encoding conformational ensembles into a single thermodynamically informed embedding improves cyclic-peptide property prediction.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_e8e8149dc2", "k": "MIRROR: Learning from the Other View for Multi-Modal Reasoning", "domain": "science", "knowledge": "Unlike large language models (LLMs) that exhibit strong reasoning capabilities, vision-language models (VLMs) struggle with visual reasoning, even on geometry problems that admit equivalent text, diagram, and combined diagram+text views. We show that these views often elicit different behaviors: a model may solve a problem from text but fail on the corresponding diagram, or succeed visually while failing textually. This inconsistency suggests that different views expose complementary reasoning paths and failure modes that standard multimodal post-training does not fully exploit. To study and exploit this phenomenon, we construct ODA-Data, a high-quality paired multimodal geometry dataset with text-dominant, image-dominant, and combined image+text views of the same problems, together with splits for training and evaluating modality-dependent reasoning behaviors. We then develop Modality-Informed Reciprocal Reasoning Optimization (MIRROR), a reinforcement learning approach for improving multimodal reasoning via self supervision. For each problem, MIRROR evaluates the model under all views, selects the best-performing view as a teacher, and trains other views with a reverse-KL objective towards the teacher. Across reasoning benchmarks that evaluate on geometry problems, MIRROR improves over standard RL and yields more accurate and consistent behavior across modalities", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_36b9e1dc42", "k": "X$^3$-OPD: Distilling Reasoning into Large Audio-Language Models via On-Policy Alignment", "domain": "science", "knowledge": "While large audio-language models have achieved remarkable progress in auditory perception, they still lag behind text-based large language models in deep logical reasoning, primarily due to the scarcity of high-quality audio reasoning data. To bridge this gap, we propose X$^3$-OPD, a cross-modal on-policy distillation framework that transfers reasoning capabilities from a powerful text teacher to an audio-language student. During training, the student generates reasoning trajectories conditioned on its own acoustic perception, while the teacher provides token-level guidance using matched textual inputs and verified answers. We further construct a three-tier symmetric corpus covering textual reasoning rendered into speech, audio-event reasoning grounded in complex acoustic scenes, and spoken-dialogue reasoning involving paralinguistic cues. This design extends cross-modal distillation beyond textually recoverable content to reasoning grounded in non-linguistic events, prosody, and conversational context. Experiments on MMSU, MMAU, BIG Bench Audio, and MMAR demonstrate that X$^3$-OPD substantially improves audio-grounded reasoning and chain-of-thought quality while largely preserving the model's existing capabilities under domain shift.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_e06800aa91", "k": "Neural solutions of coupled ghost and gluon Dyson--Schwinger equations in Landau gauge", "domain": "science", "knowledge": "The coupled ghost and gluon Dyson--Schwinger equations (DSEs) of four-dimensional Landau-gauge Yang--Mills (YM) theory are solved with a neural representation trained only from renormalized equation residuals. The neural and fixed-point solutions agree at the percent level and remain stable under changes of initialization, network size, integration grid, and infrared boundary condition. Variations of the three-gluon vertex model produce substantially larger effects than the neural error. The MiniMOM ultraviolet running and the sign change of the gluon Schwinger function are also reproduced within the limitations of the truncation.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_e2b0eed731", "k": "The Boundaries of Automation: A Theory of Persistent Human Participation", "domain": "science", "knowledge": "The rapid progress of AI has intensified the long-standing pursuit of automation: replacing human participation with algorithms wherever possible. Implicit in this pursuit is the assumption that humans remain in the loop only because current AI systems are not yet sufficiently capable. This paper challenges that assumption. Rather than asking how far automation can extend, we ask where its conceptual limits lie and argue that human participation may persist even with highly capable AI systems for three distinct reasons. Technical or complementarity grounds arise when humans contribute capabilities or perspectives unavailable to AI. Normative or developmental grounds arise when participation itself is valuable for human agency or learning. Most importantly, emergence grounds arise from target emergence: in some activities, the target is not fully specified in advance but instead emerges through the interaction itself. In these cases, human participation is not merely a means of improving execution but is constitutive of the target being produced. Human--AI co-construction, understood as the joint production of outcomes by humans and AI systems, is therefore not simply a temporary response to imperfect AI, but a persistent feature of activities whose objectives emerge through participation. This perspective has important implications for the limits of automation and for the design, evaluation, and ethics of future AI systems.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_06656f833a", "k": "Zero-Flow Two-Sample Tests", "domain": "science", "knowledge": "We propose a new approach to two-sample testing for deciding whether two sets of samples are drawn from the same distribution. The test is built on a statistical discrepancy based on the zero-flow criterion, termed zero-flow discrepancy (ZFD). We prove the validity of ZFD and propose a practical testing procedure, termed the zero-flow two-sample test (ZF2ST). The key idea is to learn how samples from the two distributions are locally misaligned and use the resulting directional pattern as evidence of distributional difference. By separating witness learning from hypothesis evaluation, ZF2ST can use flexible neural networks while maintaining valid statistical calibration. We develop both regression-based and power-maximized approaches for learning the witness. Experiments on synthetic and image datasets demonstrate that ZF2ST can achieve strong testing power for structured distributional changes while maintaining well-calibrated type-I error.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_654ded368a", "k": "Windowed-MTP: Removing the Full-Context Draft-KV Tax at Million-Token Context", "domain": "science", "knowledge": "Speculative decoding accelerates autoregressive generation by having a cheap draft propose tokens that a target verifies in parallel. Frontier models increasingly ship a built-in Multi-Token-Prediction (MTP/NEXTN) draft head under the assumption that the draft is negligibly cheap. At million-token context this breaks: an MTP draft head typically runs full attention over the entire KV cache at every draft step, so its read grows linearly with context and comes to dominate the draft cost -- precisely where speculation is most valuable. The effect compounds with draft length (a deep native draft can turn net-negative, slower than no speculation) and sharpens under hybrid/linear-attention targets, where cheaper verification leaves the draft's full-attention read exposed. We apply a StreamingLLM-style sliding window plus attention sink to the draft's attention only (Windowed-MTP), leaving full-attention verification intact. It is training-free, drop-in, and lossless by construction: the full-attention target still decides every accepted token, so windowing changes only which tokens are proposed, never which are accepted. It bounds the draft's KV working set to a constant, dropping ~99% of KV entries at 1M. Across three architecture families (Qwen GDN-MoE 35B/122B and a Mamba2-hybrid NoPE 120B) at 1M context on a single GPU in SGLang, windowing cuts the per-decode-step cost over the shipping native MTP draft by +28% to +44%, an input-invariant margin that widens with context. Since per-token latency is this cost divided by acceptance length, at matched acceptance end-to-end decode latency improves by the same amount, and more where windowing also lifts acceptance, while preserving the target's verified output distribution. Finally, the unread draft KV -- 7.7-11% of total KV at 1M -- is reclaimed via a compact ring buffer at no acceptance or quality cost.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_d33d4b72ea", "k": "Toward Generalizable Cognitive Impairment Detection with Speech-Based Multimodal Large Language Models", "domain": "science", "knowledge": "Cognitive impairment (CI) is a growing public health concern. Early and accurate diagnosis is critical for enabling timely intervention and improving patient outcomes. Speech-based CI detection has emerged as a promising non-invasive approach, as speech signals encode both linguistic and acoustic markers associated with cognitive decline. Recent advances in large language models (LLMs) further strengthen the potential of speech-based assessment by enabling more expressive representation learning and improved generalization across diverse speakers, recording devices, and clinical environments. Moreover, multimodal learning by jointly modeling linguistic and acoustic features allows for a more comprehensive characterization of cognitive and behavioral changes related to CI, leading to more reliable detection. In this work, we propose a multimodal CI detection framework based on open-source LLMs that integrates speech audio and corresponding transcripts while preserving patient privacy. Acoustic embeddings are extracted directly from speech signals, while textual embeddings are generated from automatically transcribed speech. These modality-specific embeddings are then concatenated to create a combined feature vector and used for downstream classification, without requiring access to raw or sensitive patient data. The proposed approach is evaluated on the ADReSS20 and ADReSSo21 benchmark datasets. Experimental results show that the proposed multimodal framework achieves an CI classification accuracy of 92.4% and consistently outperforms single-modality baselines. Our work establishes a new state-of-the-art for CI identification, with the proposed method demonstrating superior cross-dataset generalization. This advance highlights the power of an LLM-based multimodal framework that fuses linguistic and acoustic data to enable robust, scalable, and non-invasive screening.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_5e44aeaac6", "k": "What, Where, and How: Disentangling the Roles of Task, Language, and Model in Code Model Representations", "domain": "science", "knowledge": "Do independently trained language models come to represent the same thing in the same way? We answer for code, extending a recently introduced concept-circuit extraction method to a 2x2 design -- Python and Rust crossed with Qwen2.5-Coder-7B and DeepSeek-Coder-V1-6.7B -- and measuring a complete inventory of grammatical concepts (58 Python, 57 Rust) identically in all four cells: the smallest design that separates what depends on the task, the language, and the model. The answer splits into three parts. What earns dedicated circuitry is set by the task: the models agree on which concepts receive circuits (Spearman $ρ$ = 0.638 for Python, 0.673 for Rust, both p < $10^{-7}$). Where those circuits sit is set by the model: Qwen processes concepts in a late band (~L17-19), DeepSeek at L6-7, for both languages. How circuits grow across layers is also set by the model: Qwen gives its atomic concepts an early spike that DeepSeek does not. \"Are circuits universal?\" thus has no single answer: yes for What, no for Where and How -- universality is a property of representational content, not of computational organisation. None of this structure was fixed in advance. The agreement could have landed anywhere between independence and identity; it lands at $ρ\\approx 0.65$. Rust constructs receive 2-3x more concept-specific circuitry than their Python equivalents, in both models. Both models share neurons between the languages (6/7 and 7/7 paired constructs), DeepSeek 1.94x more than Qwen -- a direction no prior result predicts. And Qwen binds nine keywords of Rust's type-and-trait machinery into one tight neuron cluster (Jaccard 0.535 vs null 0.112, p < 0.001), a semantic dimension invisible in surface syntax. Ablation and linear probes confirm the circuits are functional. All claims are scoped to this 2x2; whether the per-model profile predicts a third model is the designed next test.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_31a102b4d6", "k": "Compact Latent Coordination for Autonomous Vehicles at Unsignalized Intersections", "domain": "science", "knowledge": "Coordinating autonomous vehicles at unsignalized intersections remains a critical challenge for multi-agent reinforcement learning (MARL) systems, which typically struggle with combinatorial action spaces, reliance on privileged information, or rigid agent designs. We propose Master-Agent Proto-plan System (MAPS), a hierarchical deep reinforcement learning (DRL) architecture in which a centralized Master agent generates a compact, continuous embedding, denoted as proto-plan, that encodes a global coordination strategy. Decentralized Worker agents integrate this embedding with local observations to execute vehicle-specific control, decoupling strategic intent from tactical execution and enabling independent optimization of each module. As a proof-of-concept evaluation of this coordination mechanism, we test MAPS across 72 intersection configurations in HighwayEnv. MAPS achieves collision-free navigation while significantly reducing average travel time, outperforming state-of-the-art baselines. The learned proto-plans further exhibit robust generalization: a system trained with three agents achieves a 94% success rate when deployed zero-shot to five-agent scenarios, confirming that proto-plan-based hierarchical learning provides a promising framework for multi-vehicle coordination.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_803fde2c36", "k": "Finite-Sample Coverage Audits for High-Recall Candidate Generation: Certification and Learning-Theoretic Design", "domain": "science", "knowledge": "An initial high-recall stage in an empirical pipeline decides which items pass to later review, labelling, or modelling, and relevant items it misses are lost to every subsequent stage. We study how many audit labels are needed to certify, with finite-sample validity, that this missed relevant mass is small, and our main results characterise the label complexity of this problem. We first show that no procedure using only labels from inside the candidate set can certify any non-trivial bound on the missed mass: the audit must sample the excluded pool, the only region where unrecovered relevant items can lie. We then prove a matching finite-corpus lower bound. Any valid audit that certifies fewer than $m$ missed relevant items with high probability when none are present, even if adaptive and permitted to label the entire included pool, must inspect on the order of $N_0/m$ excluded-pool labels. Excluded-pool auditing is therefore minimax rate-optimal, not merely convenient, for missed-mass certification in the zero-miss regime. Building on this characterisation, we develop an exact finite-sample toolkit, using binomial and hypergeometric inversion rather than asymptotic approximation, that certifies missed mass, converts it to recall through a two-pool design, certifies pre-specified families of nested candidate generators simultaneously, and produces stress-test certificates against declared perturbation mechanisms. These certificates can be paired with observable review burden to select the least burdensome pre-specified candidate generator meeting a missed-mass target. Every guarantee holds under one discipline: the candidate generator, or the pre-specified family from which it is selected, and the audit rule are fixed before the certification labels are examined.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_1701cbba98", "k": "Error Certificates for KV-Cache Eviction via Randomized Design", "domain": "science", "knowledge": "Deterministic KV-cache eviction keeps the top-$k$ tokens under an importance score and deletes the rest. We prove that this design cannot know what it destroyed: evicted values can be altered so that everything the serving system retains is unchanged while the true attention-output error grows arbitrarily, so no serving-time estimator of that error is consistent. Randomized eviction restores identifiability. With a Poisson-sampled tail at known inclusion probabilities, one logit offset performs the Hájek correction inside the softmax, and a survey-sampling variance estimator over the retained set becomes a per-step error certificate with 0.97 empirical coverage at no accuracy cost. On real workloads we pre-registered seven claims and lost three: question-aware eviction at 25--50\\% budgets is nearly free; output log-probability predicts failure better than the certificate; certificate-gated budget escalation adds nothing. What survives is attribution: the certificate separates cache-induced from inherent failures (AUC 0.73--0.75, against 0.47--0.54 for output confidence) and schedules recomputation better than random or confidence gating. Randomization buys attribution, not prediction.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_bf409ec68b", "k": "Test-Time Scaling via Error Localization", "domain": "science", "knowledge": "Scaling inference-time computation has emerged as a reliable method to improve the performance of large language models on complex reasoning and programming tasks. However, standard approaches such as independent sampling and sequential multi-turn refinement operate without token-level credit assignment, resulting in computational inefficiency, since valid reasoning prefixes are frequently discarded. In this work, we introduce Test-Time Scaling via Error Localization (TTEL), an inference-time algorithm that utilizes fixed or environment feedback to perform token-level error localization. By comparing conditional probabilities under informed feedback against a null-context baseline, TTEL isolates the step at which an error occurred. The algorithm then truncates the trajectory and branches a new generation, maximally reusing the valid prefix. Extensive evaluations demonstrate that TTEL establishes strictly dominating Pareto frontiers across sequential reasoning domains, measured by pass-at-k vs. generated-token cost. With Qwen3-8B on LiveCodeBench, TTEL attains a pass@64 of 71.0% while generating approximately half as many tokens as independent sampling (360.4k vs. 735.0k). Generalizing to math benchmarks AIME-2025 and HMMT-2025, TTEL cleanly outperforms competing test-time baselines across both Qwen3-8B and Qwen3-4B-Thinking-2507.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_d9b6168af3", "k": "KroQuant: Kronecker-Structured Block Transforms for Efficient Post-Training Quantization of Diffusion Transformers", "domain": "science", "knowledge": "Post-training quantization (PTQ) of diffusion transformers (DiTs) to W4A4 severely degrades output quality, because activations entering each linear layer contain outliers that 4-bit formats cannot represent. The standard fix applies an invertible linear transform to the activations and its inverse to the weights before quantizing both. Normalization layers between blocks force this transform to run online at every denoising step, making its inference computation cost the binding design constraint. Existing options trade quantization quality for inference cost: per-channel scaling (SmoothQuant) is computationally cheap but impacts the magnitude of the channels, which can harm quantization accuracy; fixed Hadamard transforms yield better quantization accuracy but require large block sizes that incur a high online cost; learned full-$d$ invertible transforms calibrate best but entail a prohibitive dense $d \\times d$ matrix multiplication (GEMM) per layer per step. We propose KroQuant, a PTQ method that applies a learned Kronecker-structured invertible transform to each 32-element block of the activation, storing less than half the parameters of per-channel scaling. The block-local structure runs as small tensor-core GEMMs, and on an MI350 GPU the KroQuant quantizer kernel is up to $14\\%$ faster than the SmoothQuant kernel. Offline LoRaQ weight calibration then absorbs the residual per-weight quantization error. On PixArt-$Σ$, SANA, and FLUX.1-schnell at W4A4 (MXFP4e2), KroQuant produces outputs closer to the FP reference than SVDQuant and LoRaQ on MJHQ-30K and SDCI, while preserving or improving image quality.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_f77ef9b5b5", "k": "Climate-resilient electric vehicle charging infrastructure for sustainable cities: An interpretable causal-ensemble framework for preventive maintenance and low-carbon mobility", "domain": "science", "knowledge": "Reliable electric vehicle (EV) charging infrastructure is a cornerstone of sustainable, low-carbon cities, yet urban climate stress such as extreme heat, heavy precipitation, and humidity increasingly raises equipment fault risk and undermines the resilience of urban energy and mobility services. Shifting operation from reactive repair to preventive maintenance depends on accurate, forward-looking fault-risk prediction, a task complicated by the heterogeneous time scales of physical, behavioral, contextual, and historical signals and by forecasting over a multi-week horizon. We develop FGDSE, a feature-governed dynamic stacking ensemble that forms an interpretable decision-support system for climate-resilient charging-asset management. It partitions heterogeneous signals into four feature families, assigns each to a domain expert whose inductive bias matches the data, and adds two deep temporal experts for short-term pulses and long-term degradation; a horizon-wise gating mechanism then learns adaptive weights to forecast daily fault risk over 1 to 30 days. SHAP attribution and an X-learner extend the probabilistic output into causal decision support with post-level treatment effects. On 25 months of data from 13 stations, FGDSE surpasses twelve baselines beyond the ten-day horizon, sustains about 85% macro-recall at 30 days with an AUC decay of only 3.2 points, and reveals a shift of dominance from fault history toward climate stress. It identifies extreme heat as the sole exposure whose causal effect amplifies over time, flagging roughly 30% of posts as heat-sensitive and yielding quantitative thresholds for climate-adaptive maintenance that strengthens urban mobility resilience and sustains low-carbon travel.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_a4c22a60a1", "k": "Token Budget Saturation and Mechanistic Early Detection of Reasoning Non-Convergence in Chain-of-Thought Models", "domain": "science", "knowledge": "Chain-of-thought reasoning models such as DeepSeek-R1-Distill-Qwen-7B exhibit a bimodal convergence pattern: generations either terminate within a token budget (converged) or exhaust it without reaching a conclusion (non-converged). We characterize this phenomenon empirically, showing that converged generations achieve 90.3% accuracy on AIME 1983-2024 while non-converged ones achieve only 6.6%, with an overall convergence rate of 62.0%. We then ask whether this outcome is detectable early in the thinking chain using internal model representations. Training linear probes on hidden-state activations at token positions 50-300, we find that layer-20 activations at token 150 achieve AUC 0.608 (+-0.080, 5-fold CV), reliably above chance even at token 50. Activation probes consistently outperform behavioral baselines derived from token entropy and repetition statistics. A sweep-level permutation test yields p=0.063 (100,000 permutations), consistent with a modest signal that our sample size cannot confirm at conventional thresholds. These findings suggest that convergence fate is partially encoded in intermediate representations well before the generation ends, opening a path toward early-exit inference and adaptive compute allocation.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_19c26e565b", "k": "Context-weighted Discrete Flow Matching", "domain": "science", "knowledge": "Discrete flow matching provides a flexible framework for generative modeling on discrete structures. However, the standard factorized training objective exposes the model to targets of varying difficulty, mixing well-conditioned, predictable tokens with ambiguous, high-entropy ones. We empirically demonstrate that the uncertainty over the value of each token is closely related to the density of available context in its neighborhood. Motivated by this observation, we propose a simple modification to the underlying continuous-time Markov chain (CTMC) that incorporates local context information. Our context-weighted sampler improves generation quality with negligible computational overhead, while our scaled cross-entropy loss function reweights the training signal from different tokens and reduces generative perplexity by up to 63% on OpenWebText. Moreover, our approach matches a strong semi-autoregressive block diffusion baseline in quality while retaining the ability to perform generation in any order. These results highlight the role of local context as an important factor in discrete generative modeling and show that simple context-aware modifications can significantly improve both sampling and training efficiency.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_2dfc842237", "k": "Semantic-Aware Task Clustering for Constructive and Cooperative Multi-Tasking", "domain": "science", "knowledge": "Cooperative multi-task semantic communication (CMT-SemCom) improves task execution performance by leveraging shared representations. However, as we demonstrated in [1], cooperative multi-tasking can be either constructive or destructive, depending on the semantic relationships among tasks. To ensure constructive cooperation, we propose a semantic-aware task clustering method for CMT-SemCom. We have formulated a sequential multi-stage optimization problem in which semantically aligned tasks are clustered once after a short initial training phase, and then end-to-end (E2E) joint training is conducted exclusively within the discovered groups. Specifically, the problem decomposes into two stages: (i) a semantic clustering problem leveraging hierarchical density-based spatial clustering, and (ii) an intra-cluster E2E CMT-SemCom learning problem. Simulation results demonstrate that the proposed framework effectively mitigates destructive cooperation and negative transfer, yielding accuracy gains compared to unclustered multi-tasking and individual training baselines.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_5fbb4675fa", "k": "Cautious optimism for deep parameterized quantum circuits", "domain": "science", "knowledge": "A central challenge in quantum machine learning is understanding the scaling behavior of parameterized quantum circuits (PQCs). In particular, it remains unclear how their performance on unseen data changes as the number of trainable parameters increases. Prior works have derived formal generalization guarantees for quantum models, but it is well-known that many such results do not fully characterize generalization behavior in practice. In this work, we show that gradient-based PQCs can exhibit improved performance on unseen data as model size increases, displaying the phenomenon of double descent. This contrasts with the traditional view that larger models lead to degraded generalization. We provide analytical results rigorously underpinning this behavior by leveraging add-one-in perturbation techniques and spectral properties of random matrices. We support these results with numerical experiments on re-uploading PQCs across several data sets and training set sizes, consistently observing the predicted double descent behavior. While other obstacles on the path toward practical quantum machine learning remain, our finding that deeper parameterized quantum circuits do not necessarily exhibit degraded performance provides reasons for cautious optimism.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_10b60b735f", "k": "A Diffusion-Model Subpopulation Digital Twin for Mobile Health Deployment: A Case Study on the HeartSteps Intervention", "domain": "science", "knowledge": "Mobile-health interventions increasingly use online learning and decision making algorithms to personalize when to nudge users toward healthier behavior, but a poorly designed algorithm can burden and disengage participants. New algorithm design decisions should therefore be vetted against realistic simulated users before each real-life deployment. We propose a method to develop ``JITAI-Twins'': digital twins of a target subpopulation for comparing candidate online algorithms before a just-in-time adaptive intervention (JITAI) deployment. The method builds on a conditional time-series diffusion model that is temporally consistent (future actions do not affect the generated past), and it supports repeated updating from three sources of information, in three steps: pre-training on a large observational dataset, fine-tuning on small prior intervention deployments in related populations, and inference-time calibration to the next target population from domain-scientist expertise. We validate the twin at each pre-deployment stage of the long-running HeartSteps series (v2 through v4) of physical-activity suggestion intervention deployments, treating each successive deployment as an upcoming study. The proposed method reproduces the target subpopulation's temporal and between-participant structure better than simpler simulators. These results suggest that our twin can be used to simulate a target deployment before it runs, the prerequisite for testing and informing online algorithm design decisions.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_f872602032", "k": "Mean-to-Score Discrete Diffusion: Posterior-Mean Denoisers for Score Entropy", "domain": "science", "knowledge": "Score Entropy Discrete Diffusion (SEDD) parameterizes discrete reverse processes with unconstrained positive score ratios. While positivity guarantees nonnegative reverse jump rates, it does not ensure Bayes realizability: ratios at a noisy state need not be jointly induced by any clean-token posterior under the forward kernel. The score-entropy loss has the correct population optimum but does not enforce this constraint away from it. In a trained pure-uniform SEDD checkpoint, roughly one quarter of complete score vectors violate the coordinate box, while more than half lie inside it yet remain materially incompatible with any valid posterior. Such violations can produce negative pre-normalization weights in finite-step sampling. Projecting raw scores onto the bridge polytope removes all observed negative weights and improves external generative PPL from $203.6$ to $175.1$ without changing the sampler. We introduce \\emph{mean-to-score} (M2S), which predicts a clean-token posterior mean and converts it to the score through an exact kernel-dependent linear map. The construction applies to any known coordinate-wise continuous-time Markov chain (CTMC) satisfying a mild support condition. For uniform corruption, it maps the probability simplex onto the bridge polytope; for absorbing-mask corruption, the resulting objective recovers MD4 exactly. In a controlled 28.4M-parameter CIFAR-10 comparison, M2S lowers test BPD from $3.173$ to $3.129$ and FID-50k from $\\CifarSEDDFID$ to $\\CifarMtwoSFID$. A 170M-parameter M2S model trained on about 262B OpenWebText token slots outperforms the evaluated pure-uniform SEDD, GIDD, and Neural CTMC checkpoints at every tested sampling budget, reaching generative PPL $143.3$ at 128 steps versus $183.6$ for the strongest pure-uniform baseline.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_016f1f1dba", "k": "Hilbert Operator for Progressive Encoding (HOPE): A Mathematical Framework for Deconstructing Learned Representations in Deep Networks", "domain": "science", "knowledge": "Deep neural networks encode complex representations, but deconstructing this internal knowledge remains a challenge. Given the link between learning and compression, network compression offers a promising lens to analyze this knowledge. However, standard compression heuristics often suffer from scale symmetries and architectural biases. To resolve these, we introduce Hilbert Operator for Progressive Encoding (HOPE), a mathematical framework to gradually deconstruct the representations in trained network weights. HOPE shifts network compression from the discrete domain into a Hilbert space of continuous functions. By modeling individual neurons as rank-1 Hilbert-Schmidt operators, HOPE unifies pruning and neuron merging as low-rank subspace projection. Extending this formulation, HOPE introduces macro block eviction to encompass multi-layer structures like entire residual pathways under the same unified metric. This unified approach enables unbiased architectural decisions across layers with different types and sizes. HOPE is a data-free and hyperparameter-free framework. We present proof-of-concept experiments in model compression and fine-tuning to highlight the practical potential of our theory.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_6ba55d2eb9", "k": "Emergent Misalignment Recruits a Pre-existing Persona Subspace", "domain": "science", "knowledge": "Fine-tuning an aligned language model on a narrow stream of bad advice can make it broadly misaligned on questions unrelated to the training data, a phenomenon called emergent misalignment. We ask why the narrow lesson generalizes at all, and we find that narrow fine-tuning recruits a persona structure that is present in the model before the fine-tune exists. From a frozen instruction-tuned model (Qwen2.5-14B-Instruct) we extract per-domain persona subspaces by contrastive teacher forcing and find that 4 unrelated domains share one low-rank core at 657x a random-subspace null, with 82% of that core lying outside a style core built at matched diversity. The literal first optimizer step of fine-tuning on insecure code climbs a broad-misalignment margin harder than the same code framed as educational, and forecasts realized margin movement out to 375 steps. Projecting the subspace out of the residual stream throughout fine-tuning prevents broad misalignment (27.7% to 0.0% of judged generations) while a matched-rank random subspace changes nothing; injecting it into the never-fine-tuned model induces misalignment that grows with dose to 45.4%, past the fine-tuned model it is measured against. The same projection applied to the weight gradient is inert, and three post-hoc weight edits leave the disposition in place: the sharpest edit suppresses the behavior rather than removing it, and the ablated structure re-forms inside the subspace the edit cleared. Spreading a fixed budget of bad data across 4 domains produces more broad misalignment than mechanical weight superposition and matched diversity jointly account for. All measurements come from one model at 14B; the extraction is from an aligned instruction-tuned checkpoint, which leaves the structure's provenance open; and the intervention that prevents misalignment also abolishes the narrow trained behavior.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_a19cd7d9ff", "k": "Gradient Concentration, Not Weight Saliency, Explains Representation-Level Class Unlearning", "domain": "science", "knowledge": "Machine unlearning aims to remove the influence of specific training data while preserving model utility. Many state-of-the-art approaches pursue this goal by restricting the forgetting update to a subset of parameters selected through gradient-based saliency. Although such methods are widely adopted, the actual contribution of saliency-based weight selection to representation-level forgetting remains unclear. In this work, we perform the first controlled ablation of the saliency masking mechanism used by SalUn. Using a matched-compute experimental design on CIFAR-10 and CIFAR-100 with ResNet-18, we compare saliency-based masking against random masks of equal sparsity and unconstrained updates, while keeping the unlearning objective, optimization schedule, and computational budget fixed. Across multiple representation-level evaluations, including linear probing, prototype recovery, and layer-wise CKA, the three configurations exhibit statistically equivalent representation-level recoverability. We find that forget gradients are strongly concentrated in the final network layers (approximately 92% of the squared gradient energy on CIFAR-10) before any mask is applied, causing all masking strategies to operate within the same representational subspace. Furthermore, saliency masks show limited class specificity (specificity index 0.09-0.11), selecting highly overlapping parameter subsets across different forget classes. Our findings suggest that, in the studied setting, representation-level forgetting is primarily governed by gradient concentration and representation geometry rather than by the specific identity of saliency-selected weights. More broadly, the results support a growing body of evidence indicating that effective representation-level unlearning requires objectives that act directly on latent representations rather than on increasingly sophisticated weight-selection strategies.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_be2333cc85", "k": "How Many Bits Can an Adapter Write? Measuring the Capacity and Memorization of Parameter-Efficient Fine-Tuning", "domain": "science", "knowledge": "A LoRA adapter is a few megabytes that almost everyone treats as a skill rather than a record of the data behind it. We put that assumption on a scale. Extending compression-based memorization analysis to the frozen-base setting, we measure directly, in bits, how much a low-rank adapter writes into a model it never changes. The answer is both smaller than full fine-tuning and less lawful than parameter counting would predict. Adapters store a couple of bits per trainable parameter, well short of a full model's budget, but that figure turns less on how many parameters an adapter carries than on where they sit. Move the same parameter budget from attention into the MLP and it holds nearly twice as much; strip the frozen base of its structure and the capacity all but disappears. Applied to realistic fine-tunes of Qwen2.5, the same instrument shows privacy leakage rising with the bits an adapter writes rather than the parameters it nominally has, and it draws a clean line between supervised and reinforcement learning: the secrets that supervised fine-tuning copies down verbatim, an adapter trained on verifiable rewards never records. Measuring what fine-tuning writes, rather than attacking it after the fact, turns a piece of folklore into a quantity one can design against.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_b4c692c546", "k": "M$^3$-Gen: Interpretable Multimodal Generation of Gene Expression Profiles Using Clinical and Imaging Data", "domain": "science", "knowledge": "Integrating heterogeneous biomedical data, including clinical metadata, histopathology images, and molecular profiles, is crucial for comprehensive disease understanding. However, gene expression data acquisition remains constrained by high costs and privacy concerns, limiting its use in multimodal research and AI-driven applications. We present MultiModal Molecular Generation (M$^3$-Gen), a novel framework for the generation of gene expression profiles by conditioning a Generative Adversarial Network on histopathology images and clinical metadata. M$^3$-Gen learns a unified latent representation from the clinical variables and the images, leveraging contrastive learning, and exploits the embeddings of the two modalities to guide a generative model in producing biologically coherent gene expression profiles. Evaluations on the TCGA dataset demonstrate that M$^3$-Gen generates realistic and functionally meaningful gene expression data. Importantly, by integrating multiple modalities in an attention-based mechanism, M$^3$-Gen provides intrinsic explainability: it allows the identification of which regions of the histopathology images most strongly influenced the generation of specific gene expression profiles, making the model's decisions interpretable by design.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_13f5ffb426", "k": "AI Assistants Overassist", "domain": "science", "knowledge": "Large language models (LLMs) are increasingly used as tutors and thought partners, helping users reason through problems. While guidance from AI assistants can scaffold thinking and foster learning, such benefits depend on how they help--for instance, intervening too early or too frequently may hinder true learning and cognitive engagement. Yet how AI systems navigate intervention decisions during problem-solving remains poorly understood. Here, we introduce Int-Bench, a simulation-based benchmark for evaluating LLM interventions during learning. Int-Bench simulates a \"student\" solving a problem while a \"teacher\" monitors the student's reasoning and decides whether, when, and how to intervene. Across three domains--code debugging, mathematics, and brain teasers--we evaluate LLM teachers on the frequency and timing of interventions, as well as their impact on both immediate task success and generalization to new problems. We also compare LLMs to humans, finding that LLMs intervene more frequently and earlier than humans. Moreover, in contrast to humans, they tend to provide complete solutions rather than targeted hints. These findings suggest that current LLM assistants often optimize for short-term success rather than supporting the reasoning processes needed for deeper learning and long-term success.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_d63363b922", "k": "Adaptive Depth Sparse Framework: Similarity-Driven Resource Allocation for Pre-Trained LLMs", "domain": "science", "knowledge": "Large language models (LLMs) achieve strong generation and reasoning performance, but the Transformer architecture incurs high inference cost. Existing acceleration methods often rely on task-specific fine-tuning or training from scratch, increasing adaptation cost and limiting cross-task usability. We present an Adaptive Depth Sparse Framework (AdaDSF) that converts off-the-shelf pre-trained LLMs into depth-sparse models without full retraining. Our key insight is that layers contribute unequally to representation transformation, characterized by the cosine similarity between layer input and output hidden states. Based on this, AdaDSF assigns layer-wise token retention ratios from similarity statistics, uses a lightweight router to select informative tokens at each layer, and introduces a feature-preserving alignment objective to match intermediate and final representations between sparse and dense models. On GPT-NeoX and Qwen2.5 over language modeling and commonsense reasoning, AdaDSF substantially reduces inference FLOPs while preserving performance close to dense counterparts. Under comparable sparsity, AdaDSF consistently yields smaller accuracy degradation than strong baselines including MoD, D-LLM, and DLO.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_6ea09f814e", "k": "Multi-Task Learning for Heterogeneous Prediction from Video Game State with Transfer Learning", "domain": "science", "knowledge": "Multi-task learning (MTL) is a promising approach for prediction tasks derived from video game state data, as modern game telemetry provides multiple related supervision signals from the same structured observations. We study whether a shared model trained jointly across tasks in team-based multiplayer games can improve generalization while reducing training and inference cost compared to specialized single-task models. We adapt a multimodal architecture for endpoint prediction to a general multi-task setting that combines rasterized vision inputs, global match context, and per-unit state information through an image encoder and attention-based interaction modeling. Experiments on a large proprietary World of Tanks dataset compare single-task and multi-task training, evaluate weighting strategies for mixed losses and conflicting gradients, and test pre-training/fine-tuning under limited target-data regimes. We also examine within-game transfer across game maps under structured environment shift.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_496defb5d3", "k": "The Dark Room in the Reward Channel: Dense Prediction Rewards Collapse GRPO-Trained LLM Agents -- and What Actually Works", "domain": "science", "knowledge": "Dense per-step supervision is an appealing remedy for sparse-reward, long-horizon LLM agents: reward the agent for predicting its next observation, and memory should follow. We show that under group-normalized RL (GRPO), this recipe does not merely fail -- it destroys the policy. Across Qwen3-1.7B/4B/8B on ALFWorld, a potential-based prediction reward drives every run into a degenerate absorbing state (prediction accuracy -> 1.0, task success -> 0,episode length pinned at the horizon): the \"dark room\" pathology, built automatically by the optimizer. A single-factor ablation localizes the cause -- removing only GRPO's std normalization turns the same reward from catastrophic (0%) into baseline parity -- and a two-line proposition explains why: in all-fail groups the z-scored advantage is invariant to the shaping coefficient, so bounded rewards become unbounded pressure and annealing cannot help. Our central insight generalizes this: what z-scoring amplifies is a dense signal's within-group variance while all-fail groups dominate, so signals whose variance decays by mastery are structurally amplifier-safe.This variance-profile criterion retrodicts our collapses, carries preregistered predictions for arms that had not yet run, and is consistent with published reward-channel successes (a compatibility check, not an independent test). Finally, a controlled signal-delivery matrix (identical signal, varying only the consumption mechanism) shows the reward channel is at best neutral while the auxiliary-loss channel gains ~20 points -- and a shuffled-gold placebo matches the true-gold arm, so the gap survives without correct labels. Endpoints are single-seed; seed replication and group-size controls are preregistered and in progress.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_324cc998ff", "k": "Filter Learning for Subgraphs: Algebras and Performance Risk Bounds", "domain": "science", "knowledge": "Graph signal processing tasks that leverage spectral information typically assume access to the complete graph topology, which is often unavailable in practice. We propose a systematic framework for subgraph filter learning (SFL), where subgraph-supported operators approximate ambient graph filters under partial observations. We formulate SFL as a statistical learning problem in which optimal subgraph operators are inherently data-dependent. To address the difficulty of directly estimating such operators, we develop a subgraph filter algebra based on distance-aware Laplacian constructions, defining a structured and controllable class of filters for effective approximation. We further establish performance risk bounds under the least squares loss, quantifying how well the learned operator approximates the restricted ambient mapping. Experiments real-world datasets show that, for SFL tasks, the proposed algebraic models consistently outperform polynomial filters, distribution-agnostic operators, and direct numerical filter learning baselines that attempt to recover the underlying structure from data.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_faf5e9a1a1", "k": "DART: A Degradation-Aware Recurrent Transformer for Archival Film Restoration", "domain": "science", "knowledge": "Archival film restoration is a challenging problem because historical footage contains compound degradations such as scratches, dust, blur, noise, flicker, and photometric aging, while clean reference videos are unavailable. Existing video restoration methods largely treat these degradations implicitly, reconstructing frames without explicit knowledge of where damage occurs or how severe it is. We propose DART, a degradation-aware recurrent transformer for archival film restoration. DART predicts and propagates a soft defect mask through time, using it to guide temporal fusion and condition the restoration network on both damage location and severity. This makes the restoration process explicitly aware of film artifacts rather than relying only on reconstruction losses. Experiments on real archival benchmarks show that DART improves no-reference perceptual quality over prior restoration architectures while remaining compact and efficient, producing cleaner and more temporally consistent restorations of structured film damage.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_65dff17c03", "k": "Transformer-based Diffusion models for Hydrological Time Series Probabilistic Imputation and Forecasting", "domain": "science", "knowledge": "The modeling of hydrometeorological time series with limited observations is a key challenge in the monitoring of hydro-systems and water resources, as well as for flood or drought risk assessment. Due to the high variability of the underlying processes and the sparsity of available measurements, traditional statistical approaches often struggle to accurately represent their dynamics. In this context, recent advances in deep learning offer a promising direction for improving the representation and generation of complex temporal processes sampled at several observation sites. This study investigates the application of transformer-based diffusion models to the simulation and reconstruction of hydrological time series. The proposed framework is applied to the joint modeling of water quantity and quality at six sites spread across three adjacent headwater catchments located in North-East France on a limestone plateau covered by forests and field crops. The model is calibrated and validated using available observational data, which has been quality controlled and corrected for sensor drift and malfunction through collaborative efforts by LNE metrology expertise and Andra monthly quality control over more than 15 years. Its performance is compared with several established baseline approaches commonly used for time series modeling. Quantitative evaluation metrics are employed to assess the ability of the proposed method to reproduce key temporal characteristics of the observed signals in two settings: the imputation of incomplete time series and the forecasting of upcoming hydrological conditions. Results support the effectiveness of the transformer-based approach and highlight its capacity to capture and simulate the complex patterns present in hydrological data. In particular, the results indicate that diffusion models can efficiently sample realistic time series distributions under observation settings with variable missing data for both forecasting and imputation.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_361bfb6537", "k": "Explaining Weather Bulletins via ILP", "domain": "science", "knowledge": "Inductive Logic Programming (ILP) originated within the Logic Programming community in the Nineties as a framework for combining symbolic learning with declarative knowledge representation. Nowadays, mature ILP frameworks exist and they are capable of learning complex, non-monotonic hypotheses, thus broadening both the modeling capabilities and the scope of real-world applications of ILP. This work is primarily based on the FastLAS2 framework and aims to generate simple, interpretable hypotheses to help clarify the weather bulletins issued by OSMER FVG, the Regional Meteorological Observatory of the Italian region of Friuli Venezia-Giulia. In this paper we present a pipeline that, starting from simulated meteorological raw data and from OSMERs' bulletins (used as ground truth), extracts data as ASP facts and generates ILP examples. From such examples an explanatory hypothesis is then inferred via FastLAS2. Such a hypothesis (translated into natural language) explains the weather forecast issued by human experts, and in particular the rationale behind experts' choices of specific symbols in the bulletin pictogram (the symbol-annotated meteorological map of the forecast). The proposed approach is general, not specific to any particular region and it can equally be applied to bulletins from other sources and to different regions.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_61815dd280", "k": "Automated Synthesis and Adversarial Validation of Executable Causal Research Pipelines", "domain": "science", "knowledge": "While automated research systems promise to accelerate empirical analysis, they are prone to silent failures: instances in which analysis code executes successfully yet relies on invalid causal assumptions. We present the Artificial Intelligence (AI)-based Epidemiology Research Assistant (ARA), a framework that makes these failures visible by explicitly encoding causal design principles, study-specific assumptions, and methodological constraints. ARA integrates protocol construction, synthetic data generation, and adversarial validation into a unified pipeline. The framework translates natural language research questions into structured causal protocols and executable analysis code by first constructing a protocol and then generating synthetic datasets using Structural Causal Models (SCMs) with known ground-truth effects. This synthetic-data step can also support pipeline development when access to confidential data, such as medical data, is restricted. The generated analysis is then evaluated under controlled violations of identification assumptions. We evaluate ARA on the Automated Causal Reasoning Benchmark, assessing recovery of identification strategies, causal quantities, treatment and outcome variables, and consistency between generated code and approved protocol. Protocol construction and adversarial validation did not consistently improve numerical agreement with benchmark estimates compared with standard LLM-based generation. However, they changed the failure mode: instead of silently returning causal estimates, ARA often surfaced protocol concerns, diagnostic failures, incomplete inference, or downgraded non-causal interpretations. These findings suggest that validity-first automated science systems should be evaluated not only by answer accuracy, but also by whether they indicate when causal claims are unwarranted.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_a3559acf96", "k": "Agree on the Model, Verify the Inference: GKR Protocols for HND-Based Transformer Inference", "domain": "science", "knowledge": "Outsourced Transformer inference exposes clients to model substitution and incomplete execution, while direct replay removes the computational benefit of delegation. We present GKR-HND, a registered-model protocol for verifying the polynomial backbone of Homomorphic--Nonhomomorphic Decomposition Transformers. The retained verifier checks the GKR transcript and registered-weight openings, but delegates expensive public evaluations to an assigned computation worker. Assuming an honest retained verifier and prover--worker non-collusion, the verifier accepts only when the worker's signed, request-bound response agrees with the proof claims. Experiments with pretrained HND models validate the proof path and the delegated public computation without dense-matrix replay.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_f3eac19d73", "k": "Safety-oriented sidewalk and road segmentation for smartphone-based assistive navigation", "domain": "science", "knowledge": "Independent sidewalk mobility is essential for blind and visually impaired pedestrians (BVIPs), yet smartphone-based assistive navigation requires perception models that distinguish walkable sidewalks from adjacent unsafe regions. This study presents a safety-oriented semantic segmentation framework for future mobile guidance. We introduce SENSATION-DS, a chest-height pedestrian-view dataset with 2,752 image-mask pairs and nine-class navigation-relevant taxonomy. External urban and sidewalk datasets were harmonized to this label space, and five segmentation architectures were evaluated using staged target-domain adaptation with mask-conditioned synthetic images and Segment Anything Model 2 (SAM2) pseudo-labels. Models were assessed using mean Intersection over Union (mIoU), road- and sidewalk-specific metrics, Road-as-Sidewalk Error Rate as a proxy false-safe measure, and Android Open Neural Network Exchange benchmarking. Synthetic augmentation generally improved segmentation accuracy, whereas SAM2 pseudo-labels more consistently reduced Road-as-Sidewalk errors. UPerNet-MobileNetV3 achieved the highest offline mIoU (0.715 +/- 0.006), while DeepLabV3Plus-MobileNetV3 achieved the lowest Road-as-Sidewalk Error Rate (0.079) and highest Android runtime at 512x384 (7.383 FPS). These results show that assistive sidewalk perception should be evaluated jointly by segmentation accuracy, proxy false-safe behavior, and smartphone deployment feasibility, while real-world benefit requires validation with BVIP users. This evaluation supports selecting models that balance accurate perception, conservative error behavior, and practical runtime.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_ebafb88220", "k": "Demographically-Informed Heat-Mortality Risk Curves via Risk Graph Neural Networks", "domain": "science", "knowledge": "Estimating heat-related mortality risk is a core task in environmental epidemiology, typically addressed with Distributed Lag Non-linear Models (DLNMs); interpretable exposure-response surfaces fitted to temperature-mortality time series. DLNMs are effective but ignore demographic and geographic context, despite well-established relevance to heat vulnerability. We propose Risk Graph Neural Networks (RGNNs), a hierarchical GNN encoder that uses granular census features to optimise DLNM coefficient vectors, preserving interpretable risk curve outputs while substantially improving predictive calibration. Evaluated across 10 regions of England and Wales on two unprecedented heat years, RGNN variants maintain both lower point-errors and near-nominal uncertainty coverage during the 2022 heatwave where baselines collapse.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_a7d487487b", "k": "Approximate Quantum State Preparation Through Proximal Policy Optimization", "domain": "science", "knowledge": "In this work, a quantum architecture search framework for approximate quantum state preparation (QSP) is proposed. QSP is a challenging task, since the search space grows exponentially with the number of qubits, making the identification of the optimal circuit non-trivial. To address this problem, deep reinforcement learning is employed through an agent based on proximal policy optimization. The objective of the agent is to identify the best possible approximation of the target state while simultaneously minimizing the number of gates used. At each step, the agent appends a new gate to the circuit and recomputes the fidelity between the approximated state and the target states. Various experiments have been performed from 2 to 5 qubits. Both predefined states, such as Bell, GHZ, W, and Dicke states, and completely random states are considered. The proposed framework is able to achieve approximation errors of $10^{-14}$.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_b969d39a1e", "k": "Relative Value Learning", "domain": "science", "knowledge": "In reinforcement learning, critics typically estimate absolute state values $V(s)$, estimating how good a particular situation is in isolation. However, it turns out that only differences in value are relevant for control. Motivated by this, we propose Relative Value Learning (RV), a framework that learns value differences directly via an antisymmetric function $Δ(s_i, s_j) = V(s_i) - V(s_j)$. We introduce a pairwise Bellman operator and prove it is a $γ$-contraction with a unique fixed point equal to the true value differences, derive well-posed $1$-step, $n$-step and $λ$-return targets and reconstruct generalized advantage estimation from pairwise differences to obtain an unbiased policy-gradient estimator (R-GAE). Beyond theoretical results, we integrate RV with PPO and achieve competitive performance on the Atari benchmark (49 ALE games) compared to standard PPO, indicating that relative value estimation is an effective alternative to absolute critics.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_474b79d22a", "k": "GlucoTune: A Unified Framework for Blood Glucose Preprocessing, Forecasting, and Benchmarking in Diabetes", "domain": "science", "knowledge": "Preprocessing blood glucose time-series data is a critical yet often overlooked step in developing data-driven methods for diabetes management, particularly for type 1 diabetes. The lack of standardized preprocessing workflows and evaluation protocols hinders reproducibility and complicates fair comparison across studies. These challenges are further exacerbated by data-sharing restrictions, as privacy and licensing constraints often prevent the redistribution of preprocessed medical datasets. To address these limitations, we present GlucoTune, a comprehensive and extensible framework for reproducible experimentation with blood glucose time-series data. The framework standardizes the entire experimental workflow, from preprocessing to model evaluation, enabling reproducible experiments directly from the original datasets. Reproducible preprocessing is achieved through configurable pipelines defined in portable YAML configuration files, ensuring consistent data handling without distributing sensitive preprocessed data. Beyond preprocessing, GlucoTune provides a unified interface for implementing, training, and evaluating blood glucose prediction models. The framework integrates public datasets through standardized wrappers and provides a curated collection of state-of-the-art blood glucose prediction and general time-series forecasting methods, while remaining readily extensible to additional datasets, preprocessing strategies, and forecasting models. To promote transparent and consistent evaluation, GlucoTune includes a benchmarking leaderboard that reports results across datasets, preprocessing configurations, and forecasting methods, enabling systematic comparison of experimental settings. We demonstrate the effectiveness of GlucoTune through comprehensive experiments and assess its usability in a user study.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_0109141cdf", "k": "TOUR: A Trajectory-Level Unlearning Benchmark for Offline Reinforcement Learning", "domain": "science", "knowledge": "Offline Reinforcement Learning (RL) agents are trained on fixed behavioral trajectories, which makes trajectory-level deletion important when selected data must be removed after training. Evaluating such deletion is difficult because a lower membership score can reflect trajectory removal, residual memorization visible to another attack, or policy collapse that destroys useful behavior. We introduce Trajectory-level memOrization and Unlearning in offline RL (TOUR), a benchmark that combines trajectory-level partitioning, matched non-member controls, retraining references, retained-performance anchors, and multi-attack privacy auditing. Across D4RL locomotion experiments and an exploratory AntMaze extension, TOUR shows that common deletion baselines have environment-dependent privacy-utility behavior. Retraining and fine-tuning often provide stronger retained-utility references than uniform GA+Refit, while TrajDeleter remains a useful comparator but is not uniformly stronger under the same audit. Reference-model, threshold, deviation, equivalence, action-error, representation-based, and query-limited attacks further show that a single likelihood-based membership score can overstate deletion quality. In the evaluated settings, conclusions about offline RL unlearning are therefore not stable under single-score auditing. They depend on matched non-member construction, retraining-relative calibration, attack family, retained utility, and explicit scope for diagnostic architecture or component-level evidence.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_00b864f25f", "k": "Smooth Neural Point Processes via B-Splines", "domain": "science", "knowledge": "Temporal point processes (TPPs) provide a general and flexible framework for modeling sequences of events in continuous time. Neural networks have been successfully employed to model TPPs in a highly expressive and data-driven way. Neural TPPs are typically trained via Maximum Likelihood Estimation (MLE) by minimizing the negative log-likelihood (NLL), which depends on both the conditional intensity function (CIF) and its integral over time, the compensator. Recent neural TPP approaches enable exact evaluation of the NLL without numerical integration. However, these methods typically model the compensator rather than the CIF directly, impose constraints on the neural network architecture, and are computationally expensive during training, as event contributions to the NLL are evaluated sequentially rather than in parallel. In this work, we propose a novel neural TPP model that directly parametrizes the CIF as a non-negative combination of B-spline basis functions, whose coefficients are predicted by a neural network. This formulation enables exact evaluation of the NLL, preserves full flexibility in the neural architecture, allows efficient parallelization during training, and naturally supports CIF smoothness regularization through the integrated squared second derivative. Experiments on both synthetic and real-world datasets show improved computational efficiency and predictive accuracy compared to the reference neural TPP baseline.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_4a5100e828", "k": "A Polynomial Architecture-Attribution Co-Design Framework for Exact Aumann-Shapley Attribution in GNNs", "domain": "science", "knowledge": "We study feature-level and node-level explanations for graph neural networks (GNNs) through the lens of Aumann-Shapley attribution. Path-integral methods such as Integrated Gradients provide an axiomatic formulation of attribution, but their practical use in deep GNNs typically relies on finite-sample numerical approximations to the path integral, requiring a trade-off between quadrature error and computational cost. This paper proposes APEX, a model-attribution co-design framework that makes the attribution integral exactly computable under a polynomial GNN architecture. The key component is PolyGIN, a GIN-style graph network whose message-passing, normalization, and transformation operations preserve a bounded multivariate polynomial form for scalar model scores, such as pre-softmax logits. We show that, for a PolyGIN with $L$ polynomial transformation blocks, the derivative along the attribution path has degree at most $2^L-1$. Therefore, Gauss--Legendre quadrature can evaluate the Aumann--Shapley path integral exactly, up to floating-point precision, with $2^{L-1}$ deterministic evaluation points. The resulting attributions can be computed at the feature level and then aggregated into node-level scores while preserving completeness. Experiments on synthetic and real-world graph benchmarks show that PolyGIN maintains competitive predictive performance, while the complete APEX framework achieves higher attribution fidelity than the compared baselines and substantially reduces the number of evaluations required for path integration.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_0bcef3705d", "k": "Training Large Language Models for Self-Explanation Faithfulness", "domain": "science", "knowledge": "We propose a Reinforcement Learning (RL) method to directly optimize the faithfulness of self-explanations - the extent to which a model's generated reasoning accurately reflects its internal decision-making process. While existing work focuses on evaluating faithfulness or using inference-time prompting frameworks to improve an LLM's self-explanation's tractability, these approaches do not provide a mechanism to directly optimize a model's parameters to generate faithful self-explanations. We bridge this gap by modifying existing faithfulness metrics into an RL training objective. We investigate (1) if models can be trained to accurately detect factors that affect their decisions, and (2) whether RL can directly optimize for the disclosure of these factors thereby improving LLM self-explanations' faithfulness. We experiment with two intervention types: random-word insertions and user-bias insertions, using a per-sample reward derived from the Phi-CCT correlation metric. RL fine-tuned Llama3.1-8B and Qwen3-8B show substantial improvements on the Phi-CCT faithfulness metric, with in-distribution scores rising from near-zero to as high as 0.664, and out-of-distribution scores reaching up to 0.691 on held-out tasks such as StrategyQA. Cross-intervention generalization is weaker but more interesting: a priori we would not expect a model trained only on random word insertions to generalize to user-bias phrases, yet Llama3.1-8B shows non-zero transfer in this direction. The reverse direction and Qwen3-8B do not replicate this, indicating model-dependent and setup-dependent effects we cannot yet explain. Lastly we analyze model behavior to rule out reward gaming behaviors that often plague RL training. Ultimately, we show that models can be trained to implicitly identify influential factors and disclose them, offering a scalable path toward reducing unfaithful reasoning in LLMs.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_52ed66164b", "k": "CASC: Causal Adversarial Subspace Clustering for Multivariate Spatiotemporal Data", "domain": "science", "knowledge": "Deep subspace clustering plays a critical role in applications involving multivariate spatiotemporal data, such as sea ice monitoring, disease spread analysis, and tracking neuro-degeneration over time. Despite recent advances, existing methods primarily rely on geometric self-expressiveness, assume static subspace structures, and often fail to capture causal dependencies, local spatial interactions, and long-range temporal dynamics inherent in complex spatiotemporal systems. To address these limitations, we propose a novel Causal Adversarial Subspace Clustering (CASC) framework for discovering evolving latent regimes in high-dimensional spatiotemporal data. CASC integrates a U-Net-inspired deep adversarial clustering architecture with stacked FAConvLSTM layers to preserve spatial and temporal structure while learning robust latent representations. A graph attention transformer-based self-expressive network is introduced to jointly model local spatial relationships, global dependencies, and long-range temporal interactions. Furthermore, we propose two new learning objectives: (1) a Causal Subspace Preservation Loss that aligns self-expression coefficients with latent causal relationships, encouraging clusters to reflect underlying causal processes rather than simple feature similarity, and (2) a Dynamic Temporal Subspace Evolution Loss that captures evolving subspace structures and temporal regime transitions in nonstationary environments. Together, these components transform deep subspace clustering from a correlation-driven paradigm into a causal-temporal regime discovery framework.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_ea315f49f9", "k": "Automatic knot selection in smooth additive models", "domain": "science", "knowledge": "B-spline regression constitutes a widely used framework for nonparametric modeling. The performance of this methodology depends on specifying the number and placement of changepoints, known as knots, prior to the estimation process. Such knot sequence determines the dimension of the B-spline basis used to represent the regression function and the number of coefficients to be estimated. Therefore, the knots' choice affects the model's flexibility, influencing its smoothness and goodness-of-fit. Traditionally, this problem has been addressed either by explicitly selecting knots, via knot-selection algorithms, or by regularization methods, such as P-splines, which automatically tune the regressor's smoothness. The latter have become the standard in generalized additive models (GAMs). In contrast, knot-selection techniques, frequently neglected because of computational or modeling limitations, provide certain advantages which can be valuable in some contexts. In this work, we introduce a novel explicit knot-selection technique for GAMs based on an extension of the adaptive splines (A-splines) knot selection methodology, combined with a customized Fellner-Schall scheme for tuning the associated parameters. Our approach is evaluated on various synthetic and real datasets and compared with P-splines and state-of-the-art knot-selection techniques. The results indicate comparable performance, while producing models built on a substantially smaller number of basis elements.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_5ad7b49226", "k": "Nipping the Butterfly Effect in the Bud: Self-Output Fine-Tuning for Autoregressive Weather Prediction", "domain": "science", "knowledge": "Long-horizon weather forecasting is a fundamental challenge in atmospheric science, for which autoregressive Deep Learning Weather Prediction (DLWP) has emerged as the primary paradigm. Although the autoregressive pipeline is highly scalable and flexible, its prediction errors grow rapidly over long forecasting horizons. In this work, we study this error growth phenomenon from both theoretical and empirical perspectives. Our analysis reveals that the growth is driven by a feedback loop between output errors and input distribution shifts. Specifically, the autoregressive process amplifies small initial output errors, which progressively corrupt subsequent input distributions, echoing the butterfly effect in atmospheric science and ultimately deteriorating forecasting accuracy over longer horizons. Furthermore, we show that this distributional shift originates at the earliest stage of inference, with out-of-distribution signatures detectable as early as the first autoregressive step. To mitigate this issue, we propose \\textbf{Self-Output Fine-Tuning (SOFT)}, a plug-and-play strategy that leverages the model's own one-step predictions to calibrate the biased input distribution encountered at the first step. Extensive experiments demonstrate that, despite its simplicity, SOFT achieves state-of-the-art performance on long-horizon forecasting tasks and substantially reduces both prediction errors and distributional discrepancy. The success of SOFT highlights the importance of reexamining the fundamental pipeline of deep learning weather prediction, representing a critical pipeline advance for atmospheric science.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_aad3f3b5f4", "k": "Spectral Transformation for Layer-wise Global Rank Discovery in Federated LoRA for Vision Transformers", "domain": "science", "knowledge": "Fine-tuning Vision Transformers (ViTs) with low-rank adapters (LoRA) promises better communication efficiency under federated setup, yet existing aggregation strategies face fundamental limitations. Independently averaging these LoRA factors is mathematically inconsistent, introducing cross-term aggregation error. In contrast, approaches that preserve heterogeneous client ranks by concatenating local adapters on the server substantially increase download cost and often require merging global LoRA updates into pretrained weights on the clients, causing reinitialization lag and unstable convergence. Other approaches further increase server-side overhead by reconstructing dense weight updates or training auxiliary models to refine aggregation error. In this work, we propose SpecTraL, spectral transformation for layer-wise global rank discovery, that resolves these challenges within a unified design. SpecTraL stacks local LoRA modules from clients and performs orthonormal Householder Transformation of the stacked adapters directly in the low-rank latent space, eliminating dense reconstruction of the global update and any auxiliary refinement on the server. By leveraging the Spiked Covariance Model from Random Matrix Theory, SpecTraL analytically separates the global consensus signal from non-IID noise, discovering optimal layer-wise global ranks without manual hyperparameter tuning. To match local ranks in subsequent rounds, we introduce a padding-aware initialization framework that lets clients incorporate residual LoRA dimensions without re-merging them into the pre-trained base model. Experiments on federated fine-tuning of ViT-B/16 and ViT-L/16 over DomainNet and NICO++ demonstrate improved accuracy-communication trade-offs, reduced server computation, and elimination of hyperparameter search for rank selection. Our code is available at https://github.com/DASS-Lab-Group/SpecTraL", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_07e1bd9920", "k": "From Evaluation to Optimisation: Hierarchy-Aware Training Signals for CWE Prediction in Python", "domain": "science", "knowledge": "The original ALPHA benchmark introduced a taxonomy-aware penalty for evaluating CWE-level vulnerability prediction in Python and proposed that the penalty could theoretically also serve as a training signal. This paper provides that validation. We compare three delivery mechanisms: supervised fine-tuning, a dual-head classification loss, and reinforcement learning with a dense reward derived from the normalised penalty. We find that supervised approaches consistently regress below the zero-shot baseline under distribution shift, while GRPO succeeds. Our best policy reduces the cumulative ALPHA penalty of Qwen2.5-Coder-7B on Security Hardening and Adversarial Testing (SVEN) dataset by 27.9% under greedy decoding, and by 25.5% under sampled decoding(p = 0.005, Welch's t-test), reaching statistical parity with its 4.5x larger zero-shot teacher. We conclude that the value of a hierarchical penalty as a training signal depends largely on the directness of its delivery.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_effb6f98c0", "k": "Counterfactual Explainability Framework With CycleGAN And Counterfactual-Classifier Alignnment Score for Retinal Disease Classification", "domain": "science", "knowledge": "Automated detection of vision impairing retina-based ocular conditions from fundus images is important for early screening, timely referral and reducing dependency on specialist-only assessment, for which neural network-based deep learning (DL) models have been widely utilized. However, explainability of the DL frameworks remains a major bottleneck for clinical adoption, particularly when model decisions are not linked to retinal regions that are clinically meaningful. To address this issue, this study presents CounterFundus, a novel CycleGAN-driven counterfactual explainability framework, integrating EfficientNet-B5-based retinal disease detection with visually interpretable disease-to-normal fundus image translation. For each pathological image, the counterfactual yielded by the CycleGAN generator represents an estimated healthy counterpart and the resultant difference map is utilized to localize disease-associated retinal changes. Unlike conventional post-hoc saliency methods, CounterFundus provides counterfactual explanations through visually plausible disease-to-normal retinal translation. Thereafter, to quantify the spatial agreement between counterfactual difference maps and classifier saliency, the Counterfactual-Classifier Alignment Score (CCAS) is introduced, embedding Spearman correlation, binary IoU and pointing accuracy into a single assessment protocol. To this end, EigenCAM-aligned evaluation demonstrates that the generated counterfactual explanations remain spatially consistent with classifier-relevant retinal evidence across all CCAS dimensions. Along with that, ablation studies further confirm that CCAS-filtered counterfactual augmentation improves the downstream classification performance in fundus images, establishing CounterFundus as a clinically-grounded, explainable artificially intelligence (XAI) framework for retinal disease detection.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_8f0826b351", "k": "Regularized Optimization on Grassmann Manifold: Theory, Algorithm and Applications", "domain": "science", "knowledge": "Spectral methods are among the most widely used techniques for community detection, clustering, and graph learning. Their performance, however, critically depends on the accurate estimation of the underlying spectral subspace and can deteriorate substantially in the presence of noise, outliers, or model perturbations. To address this limitation, we propose a Regularized Projection Matrix Approximation (RPMA) framework for robust estimation of rank-$K$ projection matrices. RPMA extends classical spectral projection by incorporating a regularization term, producing projection estimates that are more robust, sparse, and interpretable. We formulate the proposed model as an optimization problem on the manifold of rank-$K$ projection matrices and exploit its geometric equivalence to the Grassmann manifold. Based on this manifold characterization, we derive the first- and second-order optimality conditions, establish the local stability of the regularized leading eigenspace, and characterize the stability of the critical-point landscape under sufficiently small regularization. To efficiently solve the resulting nonconvex optimization problem, we develop a Riemannian gradient projection algorithm with backtracking line search, together with a more efficient Cayley--Sherman--Morrison--Woodbury (Cayley--SMW) gradient method that avoids repeated eigendecompositions. Extensive experiments on both synthetic and real-world datasets demonstrate that RPMA substantially improves the recovery accuracy of projection matrices and consistently outperforms conventional spectral projection methods for community detection and clustering under noisy environments.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_f51e49fe70", "k": "Weight-norm Criticality: A Mechanism for Loss Spikes Induced by the Normalization and Weight Decay", "domain": "science", "knowledge": "Most explanations of training instability focus on \\emph{learning-rate criticality}, typically characterized by the Edge of Stability, beyond which optimization becomes unstable. We argue that, in practical deep neural network training, there is an additional and often overlooked \\emph{weight-norm criticality}. This criticality is induced by the interaction between normalization (which introduces scale-invariant components) and weight decay (which persistently shrinks parameter norms). As the weight decay coefficient increases, the norms of scale-invariant weights are progressively driven toward zero. Meanwhile, the sharpness of the loss landscape increases rapidly, destabilizing the optimization dynamics and resulting in abrupt loss spikes. This perspective provides a rationale for why weight penalties can improve generalization yet cannot be made arbitrarily strong: excessive decay drives scale-invariant weight norms past a critical boundary and destabilizes training. Our work provides a new mechanistic understanding of loss spikes through the lens of \\emph{weight-norm criticality}. Moreover, \\emph{weight-norm criticality} yields testable predictions that we validate empirically in networks with scale-invariant components, providing empirical support for the proposed mechanism.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_f66f4f75f9", "k": "ADABORD: a novel AdaBoost approach for ordinal classification", "domain": "science", "knowledge": "Ordinal Classification (OC) deals with classification tasks where the classes follow a natural order. Despite the progress in OC, many existing approaches fail to fully leverage the ordinal information, treating the problem as nominal classification and thereby losing performance potential. In this work, ADABORD, an AdaBoost framework specifically designed for ordinal classification problems, is introduced. The ordinal nature of the classes is incorporated into two key components of the well-known AdaBoost algorithm: 1) the base estimator, where decision trees with the ordinal Gini splitting criterion are proposed; 2) the error function used to update sample weights at each stage and the weights of the classifier in the final ensemble model, given by the absolute ranked probability score, a measure that accounts for both the ordering and the distance between classes. ADABORD is extensively compared against seven state-of-the-art methods on the TOC-UCO repository, the largest benchmark collection for OC to date. The experimental results, supported by statistical analysis, show that ADABORD significantly outperforms competing methods, particularly on datasets with five or more classes, where the ordinal structure becomes more pronounced. Source code, along with all experimental protocols, is publicly available to ensure reproducibility and facilitate future research in OC.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_1483ffae14", "k": "The Weight of Silence: A Causal Case for Weights Over the Scratchpad in Latent Chess Reasoning", "domain": "science", "knowledge": "Latent, or silent, reasoning lets language models carry out intermediate computation in continuous vector space instead of words, and is widely assumed to function as an internal scratchpad the model actively consults during inference. Whether that assumption survives reinforcement learning has not been tested directly: existing causal analyses of latent reasoning are confined to math and logic tasks, and compare a model's reliance on its thoughts within a single checkpoint, never before and after an RL stage. We train a chess-playing model through a staged latent-reasoning curriculum followed by reinforcement learning, and find legality climbs monotonically to 61% (from a 48% pre-RL baseline) while checkmate confabulation is eliminated entirely. To locate this gain, we run a six-condition causal intervention suite on the same model before and after RL: substituting or adding matched noise to the latent thought vectors leaves performance unchanged, ablating them causes only mild degradation, and only exact-zero vectors cause collapse. This robustness gap is itself the finding: under exact-zero corruption, legality collapses to 1% pre-RL versus 9% post-RL, a gap that survives correction for testing across the full battery; milder conditions trend similarly without independently reaching significance. RL appears to add robustness to disruption, not reliance on thought content. These results push back against the field's default assumption that latent thoughts function as an actively consulted inference-time scratchpad, and instead indicate latent reasoning's principal effect here is shaping the model's parameters during training. We also demonstrate a working RL gain in chess, a domain outside the math and logic settings where multiple groups report the same latent-reasoning-plus-RL recipe failing to improve accuracy over SFT.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_8c142e7989", "k": "Best-of-Evidence: Best-of-N Selection under Partial Verification", "domain": "science", "knowledge": "BoN improves model outputs by sampling several candidates and selecting one with a proxy score, but it assumes that complete candidates can be evaluated reliably. Many vision-language tasks instead provide only partial verification: a finding, span, value, region, or relation may be checkable even when no dependable whole-response verifier exists. Moreover, the same claim may recur across candidates with opposing stances, allowing one observation to support part of the pool and contradict another. We introduce Best-of-Evidence (BoE), an inference-time selection framework that keeps the BoN candidate pool fixed, represents reusable claims with a signed candidate--factor graph, and allocates a limited budget to evidence actions that can change the final choice. BoE formalizes selection under partial verification and provides a practical score-based controller, with the zero-budget case recovering the underlying BoN decision. Theoretically, we show that residual evidence capacity limits any evidence-driven improvement and that shared factor queries can achieve an O(log K) versus Θ(K) query separation in a factor-code model. Common-ledger experiments on four medical VQA settings show that BoE can improve fixed-pool selection and rescue some BoN failures when evidence is reliable, contrastive, and decision-relevant, while also revealing the channel-quality and candidate-generation limits that prevent universal gains.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_66b058377a", "k": "An Analytically Trained Variational Surrogate for Quantum Phase Estimation on NISQ Hardware", "domain": "science", "knowledge": "Quantum Phase Estimation (QPE) is a foundational algorithm for molecular ground-state energy estimation, but its deep circuit requirements make direct hardware execution impractical on Noisy Intermediate-Scale Quantum (NISQ) devices. We present an analytically grounded variational surrogate framework in which a shallow Variational Quantum Circuit (VQC) is trained to reproduce the QPE measurement distribution without any quantum circuit simulation. The training target is computed entirely classically via the Dirichlet kernel, evaluated directly from the Full Configuration Interaction (FCI) ground-state energy, the ancilla qubit count, and the time evolution parameter, eliminating the exponentially scaling simulation bottleneck of prior surrogate approaches. We apply this framework to the hydrogen molecule (H$_2$) with a symmetry-tapered Hamiltonian, conducting a four-stage experimental investigation on IBM Quantum hardware. Stage 1 compares linear and full entangler topologies for the $R_Y$-$R_Z$-$CZ$ ansatz, with and without XpXm Dynamical Decoupling (DD), across four distributional metrics (Hellinger distance, fidelity error, total variation distance, Jensen-Shannon divergence), identifying the linear entangler as optimal. Stage 2 varies VQC layers ($p=1$ to $5$) for the linear-entangler ansatz, identifying single-layer depth as optimal under hardware noise. Stage 3 applies this configuration to the reduced $R_Y$-$CZ$ ansatz, comparing ideal and noisy simulator-trained parameters. A supplementary noise analysis at $p \\in \\{8,64\\}$ characterizes the depth-dependent interplay between circuit depth and DD effectiveness. The framework enables faithful QPE mimicry using a linearly scaling VQC, recovering the ground-state energy within the chemical accuracy threshold (1 kcal/mol), constituting a scalable, hardware-efficient paradigm for QPE-based molecular energy estimation on NISQ devices.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_a787efc428", "k": "Three-Pronged Spectral Control for Federated Parameter Efficient Fine Tuning", "domain": "science", "knowledge": "Federated parameter-efficient fine-tuning (PEFT) enables communication-efficient adaptation of large pretrained models on decentralized edge data, but it remains fragile under non-IID client heterogeneity. In low-rank adaptation (LoRA), different clients may learn locally useful but spectrally misaligned update subspaces, causing high-variance aggregation and poor global transfer. We propose TRISHUL, a spectral-control framework for robust federated PEFT. TRISHUL follows the FL no-raw-data-sharing setting but does not itself provide formal privacy guarantees. TRISHUL uses shared frozen multi-head low-rank bases to obtain algebraically exact aggregation of compact core updates, applies nuclear norm proximal shrinkage to suppress client-specific high-rank spectral components before upload, and allocates adaptation heads non-uniformly across layers using a concave water filling budget rule derived from pretrained layer capacity. Because shrinkage is performed only on small core matrices, TRISHUL adds negligible computation and no extra per-round communication over the underlying multi-head PEFT protocol. Across vision and language benchmarks, including CIFAR-100, SVHN, 20 Newsgroups, MRQA, and GLUE with LLaMA3.2-1B, TRISHUL improves convergence, stability, and final performance over federated LoRA baselines, with greater gains under stronger heterogeneity.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_0f652f0302", "k": "RadioTrace: Transmitter-Aware Diffusion for Radio Map Estimation without Deployment-Time Fine-Tuning", "domain": "science", "knowledge": "Radio map (RM) estimation aims to reconstruct the spatial distribution of wireless signal characteristics, such as received signal strength (RSS), from sparse measurements, a task that is critical for spectrum management, interference mitigation, and localization in modern wireless networks. Traditional approaches, including interpolation and deep learning, either struggle to capture complex propagation effects or require large-scale retraining for each new sampling pattern, which limits their generalization. More recently, prior-based methods have combined pre-trained generative models with measurements to reduce the need for deployment-time model fine-tuning, but they typically treat the prior as a simple regularizer and lack explicit transmitter-aware integration. In this paper, we propose RadioTrace, a novel RM estimation framework without deployment-time fine-tuning that tightly integrates sparse RSS measurements with a frozen pre-trained diffusion prior. RadioTrace incorporates transmitter (Tx) location estimation directly into the denoising loop, iteratively refining Tx coordinates based on reconstruction quality to guide the generative process. To further enhance robustness, we introduce a propagation-guided K-means initialization that mitigates poor local minima in the Tx update and provides a geometry-consistent starting point. Moreover, we provide a stochastic stability analysis for the Tx-coordinate refinement component, showing that the Tx update remains stable under perturbations induced by diffusion sampling and Tx-map relaxation. Extensive experiments demonstrate that RadioTrace achieves competitive performance with state-of-the-art learning-based methods under random sampling, and maintains strong reconstruction quality under restricted-area sampling, highlighting its adaptability, robustness, and practical relevance.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_3fbea2dd1d", "k": "Multi-turn RL with Structural and Performance Aware Rewards for CUDA Kernel Generation", "domain": "science", "knowledge": "Reinforcement Learning with Verifiable Rewards (RLVR) has emerged as a powerful technique to enhance the reasoning capacity of LLMs for optimized code generation. However, existing RLVR approaches primarily rely on outcome-based signals such as correctness and speedup, overlooking performance-critical structural properties of programs that are essential for generating optimized code. In this work, we propose CudaPerf, a reflective RL framework that incorporates both verifiable execution rewards and structural code-aware rewards derived from parallelization features (e.g., memory coalescing, occupancy, Arithmatic Intensity, and synchronization patterns). CudaPerf operates in two stages: (1) an offline pairwise ranking module that learns to distinguish strong and weak program candidates via contrastive comparisons, and (2) an online RL training phase that jointly optimizes for correctness, performance, and structural efficiency through a unified reward signal. To further enhance learning, CudaPerf utilizes iterative refinement using execution feedback enabling progressive improvement of generated candidates. We also introduce a dataset comprising 2.9k C to CUDA and 1k PyTorch to CUDA programs, each paired with diverse input configurations and multiple CUDA implementations encompassing diverse optimization strategies. CudaPerf is evaluated across multiple benchmarks comprising both C to CUDA and PyTorch to CUDA transformations. Empirical findings suggest that CudaPerf significantly outperforms strong baselines, including Qwen-3-32B (for C to CUDA) and CUDA Agent (for PyTorch to CUDA) by achieving up to 5X & 3.32X improvements in speedup, and 17% & 7% improvements in correctness, respectively.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_8eea81e99c", "k": "HierarchicalDAEW: Domain-Aware Edge-Weighted Graph Convolution with Evidential Uncertainty for Multi-Section Spatial Gene Expression Prediction from H&E Histology", "domain": "science", "knowledge": "Spatial transcriptomics assays remain costly and technically demanding, restricting transcriptome-wide profiling to specialist settings and preventing routine clinical deployment. Predicting spatially resolved gene expression from H&E histology could close this gap, yet current methods largely ignore the underlying tissue architecture and rarely quantify how their predictions can be trusted. We introduce HierarchicalDAEW, a dual-graph architecture that addresses both gaps. On the spot graph, a Domain-Aware Edge-Weighted convolutional operator learns separate projections for inter-domain, intra-domain, and boundary edges derived from Leiden clustering, allowing the model to treat tissue heterogeneity as an explicit structural signal rather than an implicit one. A second gene-level graph then fuses protein-protein interaction priors from STRING-DB with tissue-specific co-expression through learned attention gating, propagating predictions from a landmark gene set to a broader gene panel. Reliability is handled through evidential uncertainty estimation, which produces far better calibrated confidence intervals than Monte Carlo dropout under identical conditions. Across six human Visium sections spanning breast, colorectal, prostate, and cerebellar tissue, and against thirteen published baselines, HierarchicalDAEW achieves the strongest correlation with ground-truth expression, with gains that hold up under multi-seed reproducibility checks and negative controls that rule out positional shortcuts. Ablations further confirm that both the domain-aware edge typing and the hierarchical depth are necessary to this improvement, and calibrated uncertainty estimates identify low-confidence predictions for pathologist review before clinical action.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_7317ccbf92", "k": "Information-Theoretically Secure Aggregation for Lightweight Federated Learning: Resilient to Dropouts and Adversaries", "domain": "science", "knowledge": "On-device federated learning (FL) enables privacy-preserving and personalized model training on resource-constrained devices such as smartphones and IoT nodes. To reduce communication cost, sign-based methods (e.g., signSGD) transmit one-bit gradients. However, exposing gradient signs makes them vulnerable to inference attacks, while existing secure aggregation schemes are often incompatible with such methods or incur significant computational and communication overhead. We propose a lightweight and information-theoretically secure aggregation framework tailored for sign-based FL. The framework securely computes the majority vote (MV) polynomial through single-round secure multiplication, ensuring end-to-end information-theoretic security under the honest-majority assumption while revealing only the final aggregated sign to the server. To enhance efficiency and scalability, we introduce two key techniques. First, inverse-form exponent reduction halves the effective MV polynomial degree, reducing both communication and computation costs. Second, we propose single-round secure multiplication, achieving linear offline complexity and storage with only a single online communication. Together, these techniques reduce online communication by up to 99.5% and latency by up to 85.7% compared to conventional approaches. Also, by leveraging inherent MDS-code-based decoding, the framework achieves robustness against both dropouts and adversarial behaviors, yielding accuracy gains of up to 20.65% and 10.74%, respectively. Overall, the proposed framework establishes a practical foundation for large-scale, low-latency, and information-theoretically secure aggregation in sign-based FL.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_4fbd8dca05", "k": "TwistedMerge: Certified Higher-Order Diagnostics and Abstention for Model Merging", "domain": "science", "knowledge": "Model merging combines independently trained or fine-tuned models, but pairwise alignability does not imply globally consistent alignment. We formulate merging as a finite descent problem in which checkpoints are local objects, alignment maps are transitions, and cycle products are residuals. TwistedMerge is a conservative certification pipeline that separates fixed-chart averaging, synchronization-removable gauge inconsistency, a certified central obstruction on a specified comparison complex, and nonabelian holonomy. A residual is promoted to a cohomology class only after inverse-consistency, coefficient-identification, centrality, and closure tests; otherwise the method abstains and returns an ordinary or synchronized fallback. We prove a constant-edge no-go result, frozen-complex three-way and predeclared-family error-control theorems, and a refinement test for comparison-complex sensitivity. A planted neural alignment defect is removed by cycle-consistent synchronization, showing that a nonzero cycle score alone is not a higher obstruction. Controlled central systems recover the predicted non-coboundary and projective-rank behavior, while noisy estimates move from certification to abstention without false lifts on the tested controls. A trained low-rank-adapter audit shows that naive factor averaging depends on the chosen GLr representative, whereas global factor synchronization and dense-delta SVD are stable. On natural checkpoint collections, cycle residuals do not predict merge degradation and no natural central or period-index class is certified. The results position descent theory as a falsifiable certification and abstention framework.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_5a3153bc92", "k": "Machine Learning for Charge State Characterization of Isolated Double Quantum Dots", "domain": "science", "knowledge": "Scaling semiconductor quantum dot arrays toward fault-tolerant quantum computing requires efficient tuneup of spin qubits, a process that depends on the analysis of charge stability maps (CSMs) and remains largely manual. While machine learning has been widely applied to CSM analysis in reservoir-coupled devices, automated tuning in the increasingly important isolated-mode regime has received limited attention. In isolated-mode CSMs, charge transitions appear as near-vertical lines, making them well suited to compact, task-specific models. We present two convolutional neural networks with fewer than one million parameters, trained on CSMs collected from 32 silicon metal-oxide-semiconductor (SiMOS) double-quantum-dot devices measured at approximately 1 K using an automated cryogenic probing system. Sixteen devices were used for training and sixteen were held out to evaluate cross-device generalization against hand-labeled ground truth. CSMClassifier identifies charge instability and sensor artifacts, achieving 94% macro-averaged accuracy across three quality classes on 2,407 held-out images. ChargeLineNet localizes charge-transition lines and determines electron occupancy, achieving 95.3% exact line-count accuracy on 1,131 held-out images. Combined into a single pipeline, the models correctly determine electron occupancy for 93.8% of clean held-out images. Pre-training on synthetic images substantially improves label efficiency. Fine-tuning the pre-trained model on limited experimental data maintains over 90% accuracy, whereas training from scratch degrades significantly under the same conditions. Together, the two models occupy only 6.5 MB and process images in less than 60 ms on standard laboratory hardware, demonstrating a practical path toward scalable, automated characterization and tuneup of quantum-dot devices.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_269ebc3df8", "k": "Position Bias is Hidden Behind Ceiling Effects: A Permutation Diagnostic for LLM Benchmarks", "domain": "science", "knowledge": "Position bias in multiple-choice LLM evaluation is widely cited as a confound in capability comparisons, but published measurements rely on single answer-order shuffles whose results confound the bias signal with content-level noise and sampling stochasticity. I introduce inspect_permute, an open-source extension to the inspect_ai evaluation framework that runs exhaustive answer-order permutations per question and reports the chi-squared / Cramer V signature of position bias with bootstrap confidence intervals. I apply the tool across four vendors (gpt-4o-mini, claude-haiku-4-5, gemini-2.5-flash, grok-3) on five MMLU subjects, 24,000 API calls under temperature-0 generation, with falsifier predictions pre-registered via a public SHA-256 hash before half the data was observed. Position bias turns out to be statistically detectable only within a roughly 60-95% base-accuracy Goldilocks zone. Below it, processing-load dominance swamps subject-specific signal; above it, ceiling effects compress the variance below the chi-squared test resolution. Detectable cells separate into two mechanism types: monotone A-to-D decrease (processing_load, in low-tier models) and non-monotone D-drop (content_ambiguity, in a narrow capability band). Standard MMLU places every frontier-tier model above the detection band, so absence of signal there should be read as not measurable, not unbiased. Together with the ceiling-effect characterisation in arXiv:2606.26185, this work brackets the detectable region of position-bias measurement and makes the field central question askable in a verifiable form. Package, data, preregistration under MIT.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_dc2c30cd7f", "k": "Probabilistic Residual Learning for Online Recommendations", "domain": "science", "knowledge": "Modern recommender systems are typically based on deep learning (DL) models, where a dense encoder learns representations of users and items. As a result, these systems often suffer from the black-box nature and computational complexity of the underlying models, making it difficult to systematically enhance their recommendation capabilities. To address this problem, we propose Probabilistic Residual Learning (PRL), a causal Bayesian recommendation model that models the residual between ground-truth and base predictions, enabling targeted refinement of existing systems. Specifically, PRL (1) probabilistically groups users for localized residual modeling, (2) models domain-level confounders that influence user and item representations, and (3) aggregates cluster-specific residual predictions over the confounders using do-calculus. Experiments demonstrate that our plug-and-play PRL is compatible with various base deep learning recommender systems, improving their performance while automatically discovering meaningful user clusters.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_66dadf1e5b", "k": "Multilevel Graph Wavelet Compressed Sensing with Scale-Aware Neural Recovery", "domain": "science", "knowledge": "Scientific machine learning methods such as neural operators and physics-informed neural networks have advanced engineering applications and inverse problems, but their training typically requires large volumes of simulated data. This makes data preparation and model training expensive. We propose Graph Wavelet Compressed Sensing (GWCS), a learning-based framework for offline compression of graph signals by representing them as sparse, interpretable wavelet-domain representations using the spectral graph wavelet transform. The framework combines a nonparametric multilevel importance sampler, which retains high-energy wavelet coefficients within each scale for a given compression ratio, with a scale-aware graph neural network that reconstructs the signal from the sparse coefficients. We evaluate the proposed framework on synthetic approximately band-limited graph signals over random graphs and four PDE simulation datasets over meshes, which include Turbulent Radiative Layer, Viscoelastic Instability, Kolmogorov Flow, and Dynamic Stall. We compare against graph signal sampling methods and graph autoencoder baselines. Results demonstrate that the framework achieves high reconstruction fidelity and substantial data compression compared to existing benchmarks.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_9ede6d3d5a", "k": "Offline RL with Hierarchical Action Chunking", "domain": "science", "knowledge": "Offline goal-conditioned reinforcement learning (RL) holds the promise of learning general-purpose policies from static datasets. However, scaling these methods to long-horizon tasks remains a challenge due to the curse of horizon, where value estimation errors can compound through long chains of bootstrapped Bellman backups. Existing hierarchical approaches mitigate this by decomposing tasks into subgoals, yet they often rely on low-level controllers that suffer from myopic execution and biased value estimates. In this work, we propose Hierarchical Implicit Q-Chunking (HiQC), an offline goal-conditioned RL algorithm that combines high-level latent planning with low-level action chunking. By conditioning the low-level critic on temporally extended action sequences, HiQC enables unbiased k-step value backups, compressing the horizon at both the planning and execution levels. We theoretically demonstrate that this dual decomposition results in a tighter bound on value error under a bounded per-backup error model compared to standard hierarchy or flat chunking alone. Empirically, HiQC achieves the highest aggregate performance among the compared methods on the OGBench suite, with its largest gains on long-horizon navigation tasks such as humanoid-giant.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_a841a99d1b", "k": "Beyond Heavy Log Curation: Perplexity-Based APT Detection via Unsupervised, Context-Augmented Language Models", "domain": "science", "knowledge": "Advanced Persistent Threats (APTs) remain difficult to detect because only a small fraction of events in large-scale logs are attack-related, and investigation is expensive and hard to scale. Prior machine-learning approaches can reduce analyst workload, but they often rely on heavily curated training data and sophisticated preprocessing pipelines. Building and maintaining such pipelines require substantial domain expertise and engineering cost. Motivated by insights from a study of a strong APT detection baseline, we propose CAPTAIN (Context-Augmented Perplexity-based Threat Activity log detectIoN), a perplexity-based detector that leverages general, pre-trained language models with minimal, domain-agnostic preprocessing, enabling robust scoring of long, minimally processed log entries. CAPTAIN encodes recent history with an encoder model and a Q-Former-style bridge, then injects the compact context tokens into the decoder input so that perplexity reflects temporal context. To improve stability, CAPTAIN additionally applies smoothing filters to the perplexity time series. Across APT-oriented benchmarks, CAPTAIN competes with strong existing baselines and remains robust under substantially less curated inputs, that reduces the development and operational cost of advanced log preprocessing.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_6f68dddf04", "k": "Robust Asynchronous Q-Learning under Reward and State Corruption via Batching", "domain": "science", "knowledge": "Motivated by reinforcement learning in harsh environments, we consider the problem of learning an optimal policy subject to adversarially corrupted feedback. Specifically, at each time-step, an adversary can perturb both the reward and state observations of the learner following the Huber contamination model. To defend against such data corruption, we propose {\\texttt{BR-Async-Q}}: a novel, epoch-based, robust \\(Q\\)-learning algorithm built upon two key ideas: (i) partitioning the online data stream into batches to reduce variance, and (ii) constructing robust estimates of the Bellman optimality operator using such batched data. We prove a high-probability $\\ell_\\infty$ error bound for {\\texttt{BR-Async-Q}} that matches that for vanilla \\(Q\\)-learning, up to a small additive term that scales with the fraction of corrupted samples. To our knowledge, this provides the first robustness guarantee for asynchronous \\(Q\\)-learning subject to both reward and state corruption. Furthermore, when only rewards are corrupted, the dependence of our algorithm's bound on the corruption fraction is minimax optimal.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_10b7e8ab2c", "k": "New Complexity-Theoretic Frontiers of Tractability for Neural Network Training", "domain": "science", "knowledge": "In spite of the fundamental role of neural networks in contemporary machine learning research, our understanding of the computational complexity of optimally training neural networks remains incomplete even when dealing with the simplest kinds of activation functions. Indeed, while there has been a number of very recent results that establish ever-tighter lower bounds for the problem under linear and ReLU activation functions, less progress has been made towards the identification of novel polynomial-time tractable network architectures. In this article we obtain novel algorithmic upper bounds for training linear- and ReLU-activated neural networks to optimality which push the boundaries of tractability for these problems beyond the previous state of the art. In particular, for ReLU networks we establish the polynomial-time tractability of all architectures where hidden neurons have an out-degree of $1$, improving upon the previous algorithm of Arora, Basu, Mianjy and Mukherjee. On the other hand, for networks with linear activation functions we identify the first non-trivial polynomial-time solvable class of networks by obtaining an algorithm that can optimally train network architectures satisfying a novel data throughput condition.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_622b3d6308", "k": "The Geometry of Personality: Activation Steering with Jungian Cognitive Functions", "domain": "science", "knowledge": "Activation steering enables control and interpretation of LLMs, yet existing work primarily models personality through static trait frameworks such as the Big Five. We investigate whether personality can instead be represented and controlled as a set of cognitive processes using the eight Jungian Cognitive Functions. To this end, we introduce a framework comprising a Jungian evaluation protocol and a dataset of over 2,100 role-playing character narrations. Activation steering vector extraction and evaluation experiments on Llama-3.1-8B demonstrate effective monotonic control over all eight cognitive functions through activation steering. Beyond controllability, our analysis reveals that: 1. personality information is concentrated in middle transformer layers; 2. steering vectors exhibit structured geometric relationships consistent with distinctions between rational and irrational functions; 3. effective multi-dimensional steering directions cannot be recovered as linear combinations of single-function directions. These findings provide new insights into the representation of personality in LLM activation space and establish a framework for studying interpretable, effective, and multi-dimensional personality control.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_34fdaef84a", "k": "External Clustering Validation by the Homogeneity-Parsimony Trade-off", "domain": "science", "knowledge": "Scalar metrics are often used to evaluate clusterings against known classes, but they can obscure a fundamental trade-off: clusterings should be informative about class labels while avoiding unnecessary fragmentation. Here we describe normalized scores of cluster homogeneity and parsimony that quantify this trade-off. These scores build on the information bottleneck principle, modified to not reward lossy compression. We show by example and mathematical proof that our definitions of these scores have the intuitive property of varying monotonically under cluster refinement in contrast to related proposals. Extending the information-theoretic framework beyond Shannon entropies, we furthermore derive set-matching and pair-based counterparts of the homogeneity and parsimony scores. These unify commonly used evaluation criteria and show that, in the pair-based setting, the homogeneity-parsimony trade-off recovers the receiver operating characteristic of binary classifiers. We demonstrate the framework's utility for feature selection and algorithm comparison, illustrating how considering scores jointly can clarify clustering operating points and identify Pareto-optimal solutions.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_1f764441a0", "k": "Memoir: Should a Model Write to Its Memory While It Thinks?", "domain": "science", "knowledge": "Memoir combines per-sample fast memory, shared slow parameters, variable-depth latent recurrence, and a future-latent energy objective. We test its riskiest coupling: each pondering iteration may rewrite the fast tier that the same iteration reads. On procedural associative recall with key interference, we compare a coupled arm against an otherwise identical read-only pondering arm. Both arms contain 81,738 parameters, including 76,362 trainable parameters, and use matched declared forward multiply-accumulate counts, data, optimizer, schedule, and seeds. After 240 training steps across 12 seeds, coupled recall is 0.5203 with a 95 percent interval of [0.4522, 0.5883], while read-only recall is 0.6557 with [0.5953, 0.7160]. The arms are paired per seed, and the read-only lead of 0.1354 gives a paired t of 3.23 on 11 degrees of freedom with a 95 percent interval of [0.0431, 0.2277] on the difference, winning on 10 of 12 seeds. After 960 steps across 8 seeds, both arms reach 1.0000, so the measured effect is a learning-speed penalty at a fixed budget, not a demonstrated capability penalty. That longer control is ceiling limited, leaving convergence on a non-saturating task unmeasured. A predicted failure in which memory rewriting corrupts the energy signal did not occur: the energy margin grew and held. Kernel restructuring also reduced delta-rule forward time from 0.907 ms to 0.351 ms on the stated device. Code and evidence are available at https://github.com/RightNow-AI/Memoir", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_22c0ae820f", "k": "Synthetic minority data is redundant or invalid: a data-dependent validity theory and a de-biased test", "domain": "science", "knowledge": "For two decades, the standard remedy for class-imbalanced learning has been to fabricate synthetic minority examples, and the standard evidence of their validity has been a check that cannot fail: synthetic points are scored against the very data that generated them. We de-bias the check. Validity becomes a population quantity -- the probability that a synthetic point truly belongs to the minority class -- with a consistent estimator that scores synthetic points against withheld real data. Where held-out ground truth is available, the classical test underestimates true invalidity in 96-99% of method-by-imbalance-ratio cells, while the de-biased estimator tracks it closely. We prove validity is a property of the data, not the method: class overlap sets an invalidity floor no faithful generator escapes, making oversampling redundant where classes separate and invalid where they overlap. Across 91 methods, three classifiers, and datasets spanning medicine and finance -- including a generator engineered to pass the classical check -- none clears both bars: gains over the best trivial baseline are noise-thin (median below 0.01 F1, a decision threshold's reach), and most damage calibration. We release the audit as a pip-installable test and flip the burden of proof: synthetic minority data must now demonstrate, on the data at hand, both validity and information gain.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_4460dc0ef6", "k": "Toward Mechanistic Interpretability of an AI Foundation Model Fine-Tuned for Atmospheric Chemistry", "domain": "science", "knowledge": "Weather forecasting foundation models (FMs) are increasingly fine-tuned to predict air quality, offering fast global pollution forecasts at lower computational cost than conventional chemical transport models. These FMs are typically trained on reanalysis data and generate forecasts through autoregressive rollout. They do not explicitly represent governing physical or chemical processes. Therefore, high forecast skill does not reveal whether a model has learned physical mechanisms or exploits statistical regularities in its training data. Here, we present the first study of what a FM fine-tuned for atmospheric chemistry has learned by examining Microsoft's Aurora model. We impose controlled chemical perturbations on its forecasts and test them against known photochemical relationships. We then examine the internal representations that generate these forecasts. We find that Aurora captures a first-order ozone response to reactive nitrogen but does not enforce the chemical constraints that a process-based model encodes. It generates chemically inconsistent combinations of related species and relaxes localized emission features such as wildfire plumes toward background. Internally, its representations remain largely organized around the meteorology inherited during pretraining, with little structure specific to chemistry. Using sparse autoencoders, we identify internal components that causally control the chemical forecast but do not map cleanly onto individual atmospheric processes. This work provides a framework for testing whether AI forecasting systems learn atmospheric chemistry from reanalysis data. As these models are increasingly positioned to inform environmental policy decisions, we argue that composition forecasts should also be judged by their internal mechanisms rather than by benchmark skill alone.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_00a4f11a97", "k": "Emergent Compositional Skills in Mixture-of-Experts VLAs", "domain": "science", "knowledge": "We consider the problem of learning compositional robot policies end-to-end from expert demonstrations, without any pre-specified notion of task decomposition or hierarchy. We ask whether a VLA trained with a simplified Mixture-of-Experts (MoE) action head can emergently learn to decompose tasks into reusable, interpretable primitives. We find that learned experts are heavily reused across tasks and consistently correspond to qualitatively distinct low-level behaviors, suggesting that the router implicitly learns to perform high-level sequencing while experts serve as compositional primitives. Our MoE matches the task performance of a monolithic baseline while demonstrating meaningful expert specialization, a step toward modular, interpretable robot policies that emerge from data alone.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_092028be09", "k": "Memory-Computation Tradeoffs in Semi Amortized Parametric Optimization", "domain": "science", "knowledge": "Learning-enabled decision systems often use offline data or computation to reduce online compute cost. Despite the empirical success of such approaches, there is limited general understanding of how much offline information is needed to achieve a desired accuracy under a fixed online computation budget. We study this question through the lens of amortized parametric optimization: an offline phase stores a finite memory of solved problem instances, and an online phase produces a solution to a new instance by retrieving a warm start and applying $K$ steps of projected gradient descent. We analyze this setup for smooth convex parametric optimization over a compact domain, using a nonparametric predictor built from the stored offline solutions. For $μ$-strongly convex objectives, we establish matching upper and lower bounds on the memory required to guarantee $\\varepsilon$-accuracy under a fixed online iteration budget $K$. For convex objectives satisfying a $β$-growth condition ($β>2$), we obtain near-matching bounds and identify a phase transition in $K$ beyond which additional memory provides no benefit. We further provide a general proof framework that (i) explicitly quantifies the memory cost of acceleration---how much offline memory is required to achieve a prescribed speedup over the unaided online optimizer---and (ii) identifies two key quantities driving this cost: the convergence rate of the online optimizer and the Lipschitz sensitivity of the solution map to the problem parameter. Experiments on parameterized ridge regression confirm the predicted memory--computation--accuracy tradeoffs.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_bfba62ae1e", "k": "Are Diversity Metrics Measuring Diversity? A Capability-Controlled Audit of Majority-Vote Gain in LLM Ensembles", "domain": "science", "knowledge": "Majority voting over LLMs is widely assumed to benefit from diversity, and diversity measures are used to choose which models to combine. We ask whether five such measures track diversity or mainly re-express capability, auditing them as predictors of majority-vote gain over the best member across 31,900 subsets of 30 LLMs on MMLU-Pro (29 on TruthfulQA) under explicit capability controls. Three findings emerge. First, latent complementarity is ubiquitous: oracle gain is positive in 100% of subsets, yet simple voting beats the strongest member in only 9.98% of all canonical size-3 subsets (18.71% with held-out best selection); the pooled size-2-4 rate is 1.27%, partly reflecting deterministic even-size voting behavior. Second, a joint-correctness proxy (strict diversity) is nearly collinear with one minus mean accuracy (size-3 Spearman rho = +0.991 / +0.988); raw diversity-gain associations are strongly capability-entangled and, with one exception, unstable under control. Third, three linear contingency-table statistics are algebraically non-separable; after capability control, the empirically stable remainder is a modest residual pairwise co-failure association in which more shared error corresponds to lower gain. This direction is robust, but its magnitude is configuration-dependent. Joint rawspace linear regressions treating strict diversity, disagreement, and double-fault as independent predictors are rank-deficient by construction.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_c67c6b4312", "k": "GaugeQuant: Online Learning of Quantization-Optimal Bases from LLM Symmetries", "domain": "science", "knowledge": "Transformers are known to have internal continuous symmetries that leave outputs invariant, while modifying quantization. GaugeQuant leverages this in-training by introducing a LogSumExp term to the loss that breaks the symmetries, thus selecting a basis that minimizes activation outliers. A stop-gradient operator ensures that only rotation matrices are updated, yielding the language modeling objective completely unaltered. Our requires no specific calibration data, no quantization simulation, and adds negligible training overhead. With the LLaMA-2 7B model under W4A4 quantization with group size 128, perplexity drops from 8.22 to 6.73, competing with post-training methods that require frozen models and calibration datasets. Under W4A16, perplexity drops from 11.16 to 5.45. Code is available at https://github.com/MPedraBento/gauge-quant.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_625b6a986f", "k": "Self-Supervised Bio-Inspired Robotic Trajectory Planning with Obstacle Avoidance", "domain": "science", "knowledge": "Trajectory planning is a fundamental problem in robotics, requiring the generation of collision-free and efficient trajectories in a potentially complex environment. While sampling-based planners remain the dominant approach, they are often computationally expensive, particularly in high-dimensional spaces and obstacle-rich environments. Methods based on model learning offer a promising alternative, enabling efficient planning through a bounded number of forward passes through a neural trajectory planner, but commonly suffer from low sample efficiency or limited generalisation due to their reliance on exploration or expert demonstrations. This follow-up work tests our neuro-inspired self-supervised learning framework for trajectory planning that leverages forward and inverse models as the internal supervisory mechanism in an environment that contains an obstacle. Experimental results demonstrate the feasibility of the approach while revealing a tendency of our planner to exploit the learning signal provided by the forward and inverse models. To address this issue, additional training regimes and mitigation strategies are proposed and evaluated.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_1a67e2a92a", "k": "Adaptive Confidence-weighted Expansion for Trustworthy Multi-Omics Multimodal Fusion", "domain": "science", "knowledge": "Multimodal learning is a robust approach to improve predictive performance in applications such as medical prognosis. However, the clinical applicability of models that use multimodal learning is hampered by their poor performance under noisy or uninformative data streams. Present fusion approaches often lack robust mechanisms for the dynamic assessment of data quality and for the provision of a trustable confidence score on the final prediction. This dissuades their deployment in safety-critical settings. To address these limitations, we introduce Adaptive Confidence-weighted Expansion (ACE), a novel framework to enhance the trustworthiness of multimodal fusion models. ACE first enhances the multimodal space by generating new, complementary modalities from intra-modality correlations. It then employs a dual-level confidence mechanism that (1) adaptively reweighs all modalities by their reliability before fusion and (2) estimates a global trust score over the fused, final decision. To evaluate ACE, we used four challenging multi-omics datasets (BRCA, KIPAN, LGG, and ROSMAP). ACE significantly outperforms existing state-of-the-art algorithms in both classification performance and confidence calibration. Our framework provides a more stable and robust data fusion method that facilitates the use of multimodal learning in addressing high-stakes problems.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_ba61dcefe4", "k": "Pipelined Gradient Coding", "domain": "science", "knowledge": "In large-scale machine learning, distributed training commonly involves multiple workers evaluating the gradients of the model on different dataset partitions. A common challenge is the presence of straggling workers, which may significantly slow down training. Traditional gradient coding (GC) addresses this by duplicating dataset partitions across workers, allowing for the replacement of missing gradients from stragglers. However, GC requires workers to evaluate gradients on multiple dataset partitions in each step, potentially increasing overall training time. In this paper, we propose to pipeline GC, such that gradient evaluation is segmented across multiple steps and each worker evaluates gradients on just a single dataset partition per step. We develop the pipelined version for fractional repetition (FR) and cyclic repetition (CR), two representative dataset placement schemes in GC, and prove convergence guarantees for both. Through extensive simulations and experiments on cloud infrastructure, our schemes not only significantly reduce training time but also accelerate convergence compared to GC and other baselines.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_d2d167501a", "k": "Cardinality-Decomposed Loss: Matching Training Objectives to Relation Structure in Heterogeneous Recommendation Graphs", "domain": "science", "knowledge": "Graph Neural Networks trained on heterogenous bipartite graphs form a common basis in recommendation systems. These graphs often express relations that vary in cardinality, for example, user-item preferences are one-to-many and user-attribute features are one-to-one. Traditionally, a unique loss function is applied for all of the network components which is often Bayesian Personalized Ranking (BPR). While BPR works well for the recommendation task, we find that it causes attribute embeddings to collapse to near-random geometry -- a silent failure that leaves standard ranking metrics largely unaffected and therefore invisible to conventional evaluation. This in turn pollutes user node embeddings, which are shaped by both edge types simultaneously, hurting downstream tasks like personalization, segmentation, etc. Here we propose a Cardinality-Decomposed Loss (CDL) that combines both Cross Entropy (CE) and BPR to enable the model to collectively optimize for relations across cardinalities. We confirm this CE-BPR conflict by showing the two losses compete in the shared encoder's parameter space. We evaluate CDL on five datasets spanning two structural configurations -- one-to-one attributes on user nodes (MovieLens-1M, Last.fm-360K, PayPal Audience Factory, BookCrossing) and on item nodes (Yelp) -- and find that CDL consistently improves discriminability in attribute embeddings. We also show that ranking (NDCG) improves when attributes carry meaningful preference signal, but conflicts with it when the correlation is weak. We use a lambda parameter to navigate this trade-off, and a lambda-sweep reveals that dataset behavior is governed by two graph properties -- semantic alignment and topology leakage. Semantic alignment measures whether the attribute predicts preferences, while topology leakage measures whether the graph's connectivity already encodes it.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_8099267fee", "k": "LLMs Get Lost in Evolving User Intent", "domain": "science", "knowledge": "As LLMs become more capable, they are increasingly deployed as collaborative agents, taking on user-delegated tasks through iterative interaction. Yet genuine interaction is inherently dynamic: users rarely specify their intent upfront, instead disclosing, revising, and reshaping it as the conversation unfolds. Despite this, LLMs are still predominantly evaluated or trained in single-turn, fully-specified settings, leaving open a fundamental question: how well do LLMs track and act on user intent as it evolves over the course of a conversation? To study this, we introduce a framework that transforms static, single-turn tasks into dynamic multi-turn conversations in which the user's intent evolves across turns--incrementally revealed, revised, and at times redirected mid-conversation--while preserving each task's original evaluation protocol, enabling existing benchmarks to be reused as controlled testbeds without new annotation. Across multiple tasks, we surface a consistent phenomenon: strong static-setting performance does not transfer to the evolving-intent setting, with substantial drops across model families. Our findings point to a fundamental gap: today's LLMs do not yet faithfully track and act on the user's evolving intent, a capability invisible to static evaluation yet critical for future collaborative agents.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_e9c5b22c0f", "k": "Leaky Language Models: Stealing Architecture and Inference Optimizations via Per-Token Timing", "domain": "science", "knowledge": "This work presents LeakyLMs, a set of attacks that leak proprietary model, architecture, and deployment information from production language models. LeakyLMs is the first to demonstrate that key model and deployment details can be inferred using only token generation timing, even when interacting through remote APIs. LeakyLMs introduces two core attacks. The first attack targets inference optimizations and deployment strategies. For example, our attack detects whether a provider uses speculative decoding, a widely deployed inference-time optimization, and further identifies the context length of the draft model used in the pipeline. Our measurements show that Google Gemini Flash 2.5 uses speculative decoding with a draft context window of approximately 128K tokens. The second attack recovers key architectural properties, including the number of transformer layers, hidden dimension size, and number of attention heads. To achieve this, LeakyLMs builds a detailed and accurate model of token-generation timing on modern NVIDIA GPUs, characterizing how latency scales with model configuration and hardware parameters. The attack then performs a search over the architecture space using this timing model. In experiments with Llama models, the near-correct architectural configuration appears in the top-10 guesses more than 90% of the time.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_1225a9c005", "k": "Perspective Latents as an Architectural Condition for Causal Emergence in Active Inference Agents", "domain": "science", "knowledge": "A recent line of work measures causal emergence in reinforcement learning agents through Integrated Information Decomposition, reporting that $Φ_r$ grows with training and tracks reward improvement. For active inference, this raises the question of how reward-free predictive organization relates to such information-theoretic signatures. I test this within an active inference agent whose architecture separates a fast perception latent $z$ from a slow global latent $g$, where $g$ is driven by prediction error and structurally decoupled from policy gradients. In a reward-free environmental regime-switching protocol, $Φ_r$ concentrates in $g$; its aggregate magnitude is largely architectural and decreases with training. The substantive effect of learning becomes legible only at the atom-compositional level: decoupling flips sign from negative to positive and becomes regime-invariant under environmental change, while downward causation carries the regime-dependent adjustment. These results identify $g$ as the architectural locus of $Φ_r$-relevant temporal organization in an active inference agent, and argue against reading scalar $Φ_r$ as a direct index of learned integration.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_6f8d93fd06", "k": "CEDAR: Causal Edge Discovery for Autoregressive Processes", "domain": "science", "knowledge": "We propose CEDAR (Causal Edge Discovery for Autoregressive Processes), a constraint-based method for lagged causal edge discovery in sparse autoregressive time series. CEDAR screens candidate cross-variable lags using AR(1)-residualized, U-centered distance correlation, then applies two targeted conditional-independence tests per significant cross-variable lag candidate and accepts at most one lag per ordered pair. A stable MCI pruning step removes indirect edges, and optional deterministic C-nodes adjust for specified trend-like nonstationarity. In sparse regimes where few lags survive screening, CEDAR requires $O(d^2)$ CI tests after screening while retaining edge-level interpretability. CEDAR is most effective when data are scarce and variables exhibit lag-1 self-dynamics; methods with richer conditioning sets become preferable as $T$ grows or when higher-order autoregressive or simultaneous multi-lag effects are common.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_e84f742590", "k": "Attribution Markets: A Fisher-Market Formulation for Fractional Credit Assignment Between Planned Tasks and Performed Actions", "domain": "science", "knowledge": "Personal and organizational planning systems maintain two records that drift apart: what was planned (a task's effort budget) and what was done (a logged action's duration and description). Existing systems bridge them with an exclusive, all-or-nothing link that strands genuinely related but unlinked effort and reports false stalls on active goals. We formulate the bridge as a quasi-linear Fisher market: planned tasks are budget-constrained buyers, performed actions are divisible goods, and a fused text/structural/temporal signal sets each buyer's valuation. Two market instruments - a seller reserve price and a buyer cash option - yield conservation, a hard budget cap, and a provable junk filter as theorems. We extend the market with a concave completion utility discounting progress as a task nears its plan; standard convergence theory for the market's algorithm does not transfer here, resolved by a satiation-threshold fixed point with existence (Brouwer) and local uniqueness under an explicit diagonal-dominance condition, validated empirically on random and adversarial instances. A de-circularized, multi-seed benchmark - observed affinity corrupted independently of the scored ground truth - surfaces a genuine weak spot: the market's sharp, zero-entropy equilibrium is more sensitive to affinity noise than entropy-regularized optimal transport's permanently smoothed one. We resolve this with a one-parameter entropy-regularized generalization unifying the two, plus a noise-adaptive rule for its regularization strength. We report full reproducibility parameters, discuss limitations candidly, and relate the result to multi-touch attribution, optimal transport, and online Fisher-market algorithms.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_ebc639f41e", "k": "Explanation-Based Runtime Verification for Trustworthy ML-driven Optical Networks", "domain": "science", "knowledge": "Machine learning (ML) models are increasingly integrated into optical network automation frameworks to support tasks such as failure management, performance monitoring and resource allocation. In these environments, ML-driven predictions may be directly coupled with control-plane actions where incorrect decisions can immediately impact service quality, resource efficiency, and network stability. As automation levels increase, ensuring the reliability of individual decisions at deployment time becomes a critical requirement. Explainable artificial intelligence (XAI) techniques have emerged to improve transparency by highlighting the factors influencing ML predictions. In addition to identifying influential features, they provide insights into the underlying reasoning process of the model, revealing how different input variables contribute to the final outcome and how feature interactions shape the decision boundary. In this work, we introduce explanation-based runtime verification, an approach that exploits model explanations to assess the soundness of individual ML decisions before they are executed in the network control loop. The proposed approach evaluates explanation coherence and physics grounding consistency at runtime, enabling the system to defer or reject decisions flagged as uncertain. We demonstrate the effectiveness of our approach on a representative use case of lightpath quality of transmission classification. Experimental results show that explanation-based verification can intercept a significant fraction of erroneous decisions while preserving high automation rate.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_ca00596307", "k": "End-to-End Learning of Safe Optimal Feedback Control in High Dimensions with Control Barrier Function Layers", "domain": "science", "knowledge": "We consider the problem of learning high-dimensional semi-global feedback controllers under hard safety constraints enforced by control barrier functions (CBFs). Incorporating CBFs into end-to-end policy training requires embedding a quadratic-program-based safety filter as an optimization layer, but computational and differentiation bottlenecks have largely restricted prior approaches to low-dimensional systems, typically with at most 16 state dimensions. We address this limitation by combining operator splitting with the recently developed Jacobian-Free Backpropagation (JFB) method to enable scalable end-to-end training while preserving hard safety guarantees through the CBF safety filter. We justify this training methodology theoretically using nonsmooth analysis techniques and demonstrate its effectiveness on high-dimensional multi-agent nonlinear control problems with state and control dimensions up to 1200 and 400, respectively.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_a1696a8538", "k": "Cross-Domain Generalization in Optical Networks via Joint Contrastive and Classification Learning", "domain": "science", "knowledge": "The robustness of machine learning techniques across heterogeneous network domains remains an open challenge in optical networks. Models trained on data from a specific topology or operational configuration often exhibit degraded performance when deployed in unseen networks. In this work, we address this challenge by proposing a representation learning technique aimed at capturing task-relevant relationships that remain stable across domains. The proposed technique is based on a novel joint contrastive and classification learning approach in which representation learning and task optimization are performed simultaneously, allowing both objectives to shape the latent space. Experimental results on a representative use case, namely, lightpath quality of transmission estimation, demonstrate the effectiveness of our approach compared to baseline approaches, and highlight its capacity for rapid adaptation, providing excellent performance even with limited fine-tuning.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_210d282886", "k": "Writhe-Based Polymer Link Classification Using Machine Learning", "domain": "science", "knowledge": "Unique and rapid classification of knots and links is an open mathematical problem that is relevant to a range of (bio)physical systems, including polymer melts, DNA, and proteins. In this paper, we explore a data-driven approach to the classification problem of link topology. Extending the framework introduced in Ref. 1 (Sleiman et al, 2024 Soft Matter, 20(1), pp.71-78), we show that a feedforward neural network trained on the writhe density matrix classifies thermally equilibrated configurations of the first six prime links with 97% accuracy. We demonstrate that this accuracy remains high across a range of temperatures and lengths of link components, while rapidly deteriorating with the addition of topology-altering Gaussian noise; a result consistent with the writhe density matrix containing features sensitive to topology. Our results show that neural networks based on the writhe density matrix efficiently classify two-component links, establishing machine learning as a promising tool for rapid classification of more complex link topologies, e.g. Borromean rings and multi-component links, as the computational cost of exact numerical calculation of topological invariants becomes prohibitive.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_a448d89bab", "k": "Adaptive Multi-Horizon Reinforcement Learning", "domain": "science", "knowledge": "Effective decision-making in complex and changing environments requires balancing short-term and long-term consequences. In reinforcement learning (RL), this trade-off is typically controlled through a fixed discount factor, which imposes a single exponentially discounted temporal horizon. However, biological agents exhibit flexible and adaptive temporal discounting, suggesting that effective planning requires multiple timescales. Here, we propose a multi-horizon approach that adaptively selects and combines temporal horizons, enabling robust adaptation to changes in reward structure without manual discount-factor tuning. This flexibility makes the method particularly suitable for continual learning scenarios involving task switches and varying environmental configurations. Empirically, we demonstrate that our approach identifies effective discount factors across a range of MiniGrid environments, including continual settings composed of three sequentially changing tasks. These results suggest that adaptive temporal discounting can improve parameter efficiency and enhance adaptability in both artificial and biologically inspired learning systems.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_f30eff0e47", "k": "SalesLoop: Reinforcement Learning from Performance Feedback for Sales Lead Ranking", "domain": "science", "knowledge": "Lead ranking in Customer Relationship Management (CRM) systems faces a persistent challenge: models achieving high offline accuracy often underperform in production. We identify three fundamental gaps responsible for this disconnect: offline-online metric mismatch, pointwise-listwise objective misalignment, and temporal distribution drift. To address these gaps, we propose SalesLoop, a reinforcement learning framework that establishes a closed feedback loop between model predictions and real-world business outcomes. Our approach introduces (1) a performance-aware reward that encodes conversion outcomes weighted by ranking position and conversion velocity, and (2) Discriminative GRPO, a listwise optimization objective that adapts Group Relative Policy Optimization to discriminative ranking models. SalesLoop improves NDCG@K by +7.9\\% and P@K by +15.8\\% over the strongest static baseline. A 160-day production A/B test at a New Energy Vehicle manufacturer, spanning 16.5M leads and 280 sales specialists across two provincial markets, validates statistically significant cumulative lift of +4.7\\% ($p=0.047$) and +8.7\\% ($p=0.002$). In production, the ranking backbone achieves Top-10\\% recall of 44.1\\% and surfaces high-intent leads at $2.3\\times$ the conversion rate of specialist baselines.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_a1da8f1e8d", "k": "The Frankl--Tokushige product conjectures for $r$-cross-intersecting families", "domain": "science", "knowledge": "We settle the uniform and biased product conjectures of Frankl and Tokushige for $r$-cross-intersecting families. Let $r\\geq2$, let $0\\leq k_i\\leq(r-1)n/r$, and let $\\mathcal{F}_i\\subseteq\\binom{[n]}{k_i}$ be $r$-cross-intersecting. We prove the sharp inequality $$\\prod_{i=1}^r\\frac{|\\mathcal{F}_i|}{\\binom{n}{k_i}}\\leq \\prod_{i=1}^r\\frac{k_i}{n},$$ with equality attained by the corresponding levels of a common $1$-star. As a consequence, we obtain the analogous $p_i$-biased measure theorem for $0\\leq p_i\\leq(r-1)/r$, $$ \\prod_{i=1}^rμ_{p_i}(\\mathcal{F}_i)\\leq \\prod_{i=1}^r p_i.$$The main difficulty is that unequal parameters do not determine a single common target level; instead, the target levels $\\ell_1,\\ldots,\\ell_r$ must satisfy $\\sum_{i=1}^r \\ell_i=(r-1)n$. We overcome this asymmetry in three steps. An ordered-partition coupling gives a sharp additive inequality for every such choice of target levels. A star-calibrated upper-shadow inequality relates the density of a family on its original level to the density of its upper shadow on a suitably chosen target level; it is proved by induction on $n$, with the induction step reduced to a two-point inequality. Finally, an analytic inequality shows that the resulting asymmetric additive estimate implies the required product bound. Perhaps surprisingly, the coupling captures all the combinatorial information of cross-intersection, reducing the remainder of the proof to an analytic argument.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_f1e29949ac", "k": "Tight Hamilton Cycles in Linearly Quasirandom 3-Graphs", "domain": "science", "knowledge": "We study tight Hamilton cycles in linearly quasirandom 3-graphs. An $n$-vertex 3-graph $H$ is $(p,μ)$-dense if $e_H(X,Y,Z)\\ge p|X||Y||Z|-μn^3$ for all $X,Y,Z\\subseteq V(H)$. Ara{ú}jo, Piga and Schacht asked whether $p,α>1/4$ together with $δ_2(H)\\geαn$ force a tight Hamilton cycle. We give a negative answer: for every $\\varepsilon,μ>0$ and all sufficiently large $n$, there exists an $n$-vertex $(p_0-\\varepsilon,μ)$-dense 3-graph $H$ with $δ_2(H)\\ge(p_0-\\varepsilon)n$ and no tight Hamilton cycle, where $p_0:=\\max_{0\\le x\\le1}\\min\\{x^3,1-x\\}\\approx0.317672$. For every $p>1/3$, we determine the asymptotically sharp minimum-codegree threshold. Writing \\[ δ_0(p)= \\left(\\frac{1-\\sqrt{(4p-1)/3}}{2}\\right)^2, \\] we prove that every sufficiently large $(p,μ)$-dense 3-graph $H$ with $δ_2(H)\\geαn$ contains a tight Hamilton cycle whenever $α>δ_0(p)$ and $μ$ is sufficiently small. A matching construction shows that this threshold is best possible. The proof uses absorption together with a new fixed-length connecting lemma based on a regular slice, a directed pair-state graph, and a finite scalar lemma.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_86fb0f3b6a", "k": "Arithmetic progressions in a random set on a budget", "domain": "science", "knowledge": "A restricted-budget version of the random graph process, introduced by Frieze, Krivelevich, and Michaeli in 2025, studies the construction of structures by an online player who can purchase only a limited number of random edges. In this paper, we transfer this framework from random graphs to random subsets of integers, focusing on the construction of $k$-term arithmetic progressions. A player, Builder, is presented with a sequence of $t$ integers drawn uniformly at random from $[n]$. As the elements are revealed one by one, Builder must immediately and irrevocably decide whether to select the current integer, subject to a maximum budget of $b$ selected elements in total. We establish the optimal thresholds for this process, proving that for $t = ω(n^{1-2/k})$, a budget of $b = Θ((n/t)^{\\frac{k-2}{2}})$ is both necessary and sufficient for Builder to successfully construct a $k$-term arithmetic progression with high probability.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_9d3a50eb01", "k": "A new recursive construction for large sets of Kirkman triple systems", "domain": "science", "knowledge": "We give a recursive construction of an LKTS(7q + 2) from an LKTS(q + 2) and an auxiliary permutation structure on an abelian group of order q, which we call a cubic orthomorphism. We prove that cubic orthomorphisms exist whenever q is a product of prime powers congruent to 1 modulo 6. The construction yields, in particular, LKTSs of orders 93, 177, 219, 261, and 303.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_b86f20bd17", "k": "Improved lower bounds for the Shannon capacity of odd cycles", "domain": "science", "knowledge": "The Shannon capacity $Θ(G)$ of a graph $G$ quantifies the maximum rate at which information can be transmitted with zero error over a noisy channel. It is lower bounded by $α(G^d)^{1/d}$ for any $d$, where $α(G^d)$ is the independence number of the $d$-th strong power of $G$. We construct independent sets of size $134753$ in $C_7^{10}$, $21909$ in $C_{11}^{6}$, and $62530$ in $C_{13}^{6}$, improving the best known lower bounds for the Shannon capacity of these graphs to $Θ(C_7)\\geq 134753^{1/10}>3.258020$, $Θ(C_{11})\\geq 21909^{1/6}>5.289773$, and $Θ(C_{13})\\geq 62530^{1/6}>6.300109$. We also improve the best known lower bounds on the independence numbers of several individual strong powers of odd cycles that do not improve the Shannon capacity lower bound. The constructions were discovered through iterative interactions with a Large Language Model (LLM), illustrating the potential of LLMs for finding explicit combinatorial constructions.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_ebca6053f0", "k": "Interval Garside groups arising from involutions in finite reflection groups", "domain": "science", "knowledge": "We identify and study the interval Garside groups arising from the restriction of the absolute order on a Coxeter group to a lattice $[1,w]_T$, where $w$ is an involution. Those involutions $w$ for which $[1,w]_T$ is a lattice were previously classified by the second author; every such involution lies in the center of the parabolic subgroup generated by $[1,w]_T$. Except in type $B_n$, the obtained groups are isomorphic to (decomposable) right-angled Artin groups. We also investigate the situation for some finite complex reflection groups, mostly in rank two, taking for $w$ a (not necessarily involutive) central element.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_389d9c8f59", "k": "Chromatic symmetric functions of claw-free graphs are not Schur positive", "domain": "science", "knowledge": "Chromatic symmetric functions are well-studied symmetric functions in algebraic combinatorics that generalize chromatic polynomials of graphs. In 1995, Stanley introduced these symmetric functions and conjectured that they are Schur positive for claw-free graphs. We give examples of a line graphs, which are thus claw-free, whose chromatic symmetric function have a negative coefficient in its Schur expansion. We also give a counterexample to the 2018 conjecture of Monical that Schur positive chromatic symmetric functions have saturated Newton polytope when expanded in any finite number of variables. Both of these examples were found using ChatGPT-5.6 Sol Pro.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_72fd9fa00a", "k": "Adjacency-degree algebras and spectral determination of graphs", "domain": "science", "knowledge": "McKay proved that the spectra of all polynomial functions of the adjacency matrix $A$ and the diagonal degree matrix $D$ determine a tree. We prove a principal version of this theorem. Let $\\mathcal A(G)=\\langle I,A_G,D_G\\rangle$ and let $M_G=\\mathcal A(G)\\mathbf1$ be the cyclic module generated by the all-ones vector. For connected graphs the ideal $\\mathcal A(G)J\\mathcal A(G)$, where $J=\\mathbf1\\mathbf1^T$, acts on $M_G$ as the full endomorphism algebra. We show that every forest satisfies $M_G=U_G$, the automorphism-orbit module, and that the induced algebra on the orbit quotient of a tree is a full matrix algebra. It follows that the scalar moments $\\mathbf1^Tw(A_T,D_T)\\mathbf1$ determine every tree. For general graphs these moments are degree-decorated caterpillar homomorphism counts. The resulting moment-rigidity class lies inside the amenable, compact, refinable hierarchy of color refinement, and its first small-order failures are ten-vertex integral switchings invisible to $M_G$.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_47b04f6749", "k": "Symmetries of (3, 6)-Fullerenes", "domain": "science", "knowledge": "A (3, 6)-fullerene is a cubic planar graph whose faces all have 3 or 6 sides. We give an exact enumeration of (3, 6)-fullerenes with V vertices and each of five possible symmetry types: *332, 332, 2*2, *222, and 222, in orbifold notation. We use this enumeration, together with some constructions, to resolve three conjectures about the existence of (3, 6)-fullerenes with 2*2$, *222, and 222 symmetry for given numbers of vertices.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_5e3faf665d", "k": "Fatness and Flatness", "domain": "science", "knowledge": "Fat minors are the metric analog of graph minors that are tailored to the analysis of metric (edge-weighted) graphs and, more generally, metric spaces having a suitable notion of shortest paths. Despite a large interest in this notion, not much is known about the structure of metric graphs excluding a fixed fat minor. We prove that if a metric graph $G$ excludes a fixed graph $H$ as a $δ$-fat minor, for some $δ>0$, then $G$ enjoys the metric analog of flatness (aka uniform quasi-wideness) - a structural property from the field of Sparsity. In essence, our flatness result says that for any $α\\geq β$ large enough compared to $δ$, in every large enough set $A$ in $G$ one can find a sizable subset $B$ that becomes $α$-scattered after removing a bounded number of balls of radius $β$. We call this property drill-flatness. Notably, the proof only relies on excluding shallow fat minors: every branch set has radius at most $2α$. As a corollary, we prove that metric graphs that exclude a fixed $δ$-fat minor have bounded $\\varepsilon$-scatter dimension if we consider only $\\varepsilon$-scatters at distances large enough compared to $δ$. By combining this with the results of Abbasi et al. [FOCS 2023], we infer that the $k$-Center problem on instances excluding $H$ as a $δ$-fat minor admits an approximation algorithm that finds a solution of cost at most $(1+\\varepsilon)\\cdot\\mathsf{OPT}+{\\cal O}(δ/\\varepsilon^2)$ in time ${\\cal O}_{H,\\varepsilon}(n^{{\\cal O}(1)})$. This is one of the first algorithmic results for general fat-minor-free metrics. We also study drill-flatness in hereditary classes of (unweighted) graphs, where we obtain a characterization equating drill-flatness with excluding shallow induced minors. This is an induced analog of the equivalence between flatness and nowhere denseness - one of central results of Sparsity.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_f6bc08c736", "k": "A lower bound on the growth rate of $(132,213)$-avoiding cyclic permutations", "domain": "science", "knowledge": "We construct a new reduction process which takes a $(132,213)$-avoiding permutation to a shorter one that is cyclic if and only if the original was. Iterating it determines whether a given $(132,213)$-avoiding permutation is cyclic. Reversing it gives four moves that build every cyclic $(132,213)$-avoiding permutation, uniquely, from $1$ if $n$ is odd, and $21$ if $n$ is even. Our main application is the first non-trivial lower bound for the growth rate of $\\mathcal{C}_n(132,213)$, the cyclic permutations of length $n$ avoiding $132$ and $213$. We also give several other consequences of the reduction, including a bijection between the odd and even size classes and an exact enumeration for those permutations with a restricted number of layers.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_b0295e7b35", "k": "Homotopy types of intervals in corank-three higher Bruhat orders", "domain": "science", "knowledge": "We prove Reiner's conjecture for higher Bruhat orders in corank 3: the facial intervals of B(n,n-3) are precisely the spherical intervals, and all other intervals are contractible.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_f5af5c4fc2", "k": "Selection-structure generalizations of the Borsuk-Ulam theorem", "domain": "science", "knowledge": "We prove Borsuk-Ulam-type results governed by selection structures. Selection structures extend the matroidal framework for colorful theorems in discrete geometry and include non-matroidal examples such as chessboard complexes. Motivated by Frick and Wellner's Radon-type strengthening of Fan's theorem and its colorful variants, we prove selection-structure analogues whose conclusions are determined by Radon partitions. We also prove a prime-power selection-structure covering version of Volovikov's theorem, governed by Tverberg partitions. We include applications to fair partitions, including selection-structure versions of the ham sandwich and necklace splitting theorems.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_fc6ea8e177", "k": "A six-neuron counterexample to the target-free clique conjecture", "domain": "science", "knowledge": "The target-free clique conjecture asserts that the supports of stable fixed points of a nondegenerate combinatorial threshold-linear network (CTLN) are exactly its target-free cliques: bidirected cliques for which no outside vertex receives an edge from every clique vertex. We give an explicit six-neuron counterexample. For every sufficiently small $\\varepsilon>0$, the CTLN defined by one fixed graph at $δ=29\\varepsilon/25$ is nondegenerate and has a stable fixed point with nonclique full support. Its values of $q=δ(1-\\varepsilon)/\\varepsilon$ tend to $29/25$. In the complementary direction, for any CTLN on $n\\geq3$ vertices, we prove that in the parameter range \\[ q\\geq n-2-\\frac{n-3}{2}\\varepsilon, \\] no nonclique support can satisfy both the fixed-point positivity and linear stability conditions. Consequently, throughout this range, every nondegenerate CTLN has exactly its target-free cliques as supports of stable fixed points. In particular, this holds when $\\varepsilon\\leqδ/(δ+n-2)$.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_3b953ad0ab", "k": "A Weighted Sum Formula for Double Eisenstein Series", "domain": "science", "knowledge": "We prove a weighted sum formula for double Eisenstein series. Its corresponding identity for the generating series of multiple divisor sums was conjectured by the author in his master's thesis. The double Eisenstein series identity follows from the restricted double-shuffle relations proved by the author and Tasaka, while the proof of the divisor-sum identity is combinatorial and uses generating series.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_b2d3f996d2", "k": "Frieze patterns and aperiodic tilings of the plane", "domain": "science", "knowledge": "This short note provides two examples of aperiodic frieze patterns of the plane, supported on the rhombic Penrose tiling and the Godrèche--Lançon--Billard tiling. That is, we provide a decoration of their vertices with positive integers which satisfy the diamond rule, in analogy to the usual (in)finite frieze patterns as defined by Conway and Coxeter.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_b35165b299", "k": "The Ehrhart series of magic squares of order seven", "domain": "science", "knowledge": "Let $\\mathrm{IMS}_n(m)$ denote the number of $n\\times n$ nonnegative integer matrices whose row sums, column sums, and two main diagonal sums are all equal to $m$. We determine the Ehrhart series $F_7(q)=\\sum_{m\\ge 0}\\mathrm{IMS}_7(m)q^m$ as a reduced rational function. The denominator has degree $373$ and cyclotomic factors of order at most $15$; the numerator is a palindromic polynomial of degree $366$ with nonnegative integer coefficients. Using the SimpCone decomposition, the associated polytope is represented as a sum of $166$ million signed simplicial cones. The LRQC evaluator computes their generating functions over finite fields; a typical cone requires only one or two quotient characters, and the cost per character is nearly linear in the truncation degree $T$. An explicit common denominator together with Ehrhart reciprocity reduces the rational reconstruction to the prefix up to $T=1256$, while an explicit counting bound supplies the coefficient bounds needed for deterministic lifting from the prime fields to $\\mathbb Z$. This prefix is independently computed for the whole cone family in eight prime fields. Exact Chinese remaindering lifts the verified residues to equality over $\\mathbb Z$, the finite-prefix criterion proves the rational identity, and exact polynomial gcds prove that the displayed denominator is reduced.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_a60ae08062", "k": "Critical-exponent stratification and inverse realization on biregular trees", "domain": "science", "knowledge": "For free type-preserving discrete actions on the biregular tree $\\mathcal T_{r+1,s+1}$, we stratify the critical-exponent spectrum by quotient complexity. The unrestricted spectrum is the full interval $[0,\\frac12\\log(rs)]$, whereas the finitely generated spectrum is countable and dense and is encoded by the Hashimoto radii of finite typed cores. At fixed rank, finitely many typed kernels parametrize all values, and every nonzero accumulation belongs to a lower-rank stratum. Rank two admits a complete effective inverse classification through the figure-eight, theta, and dumbbell polynomial families.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_1a028b6855", "k": "An $O(t\\log^2 t)$ Bound for $k$-Connected Subgraphs in Dense $K_t$-Minor-Free Graphs", "domain": "science", "knowledge": "Delcourt and Postle reduced the Linear Hadwiger Conjecture to coloring $K_t$-minor-free graphs on $O(t\\log^4 t)$ vertices. An important theorem in their proof process asserts that every sufficiently dense $K_t$-minor-free graph contains a small, highly connected subgraph. In this paper, we show that such a subgraph can be chosen to be smaller. More precisely, there exists an integer constant $C \\ge 1$ such that, for all integers $t \\ge 3$ and $k \\ge t$, every $K_t$-minor-free graph $G$ with $d(G) \\ge Ck$ contains a nonempty $k$-connected subgraph $H$ satisfying $v(H) \\le C^2 t\\log^2 t$. Thus the structural bound improves from $O(t\\log^3 t)$ to $O(t\\log^2 t)$, and the graphs occurring in the reduction have order $O(t\\log^3 t)$ rather than $O(t\\log^4 t)$.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_295f907861", "k": "Skeleton Chordalities", "domain": "science", "knowledge": "We study new higher-dimensional analogs of graph chordality and review the existing ones. Our main results for simplicial complexes are: (1) $Δ$ skeleton-E-chordal $\\Rightarrow$ $Δ^\\vee$ vertex-decomposable $\\Rightarrow$ $Δ$ skeleton-clique-chordal. Moreover, for subflag complexes, $Δ$ skeleton-E-chordal $\\Longleftrightarrow$ $Δ^\\vee$ vertex-decomposable. (For $d=1$ this boils down to ``$G$ chordal $\\Longleftrightarrow$ $G^\\vee$ vertex-decomposable'', a result closely related to Fröberg's theorem.) (2) For subflag complexes, $Δ$ is skeleton-E-chordal $\\Longleftrightarrow$ it splits as $Δ= Δ_1 \\cup Δ_2$, with each $Δ_i$ a skeleton-E-chordal induced subcomplex of $Δ$, and with $Δ_1 \\cap Δ_2$ a complex whose $1$-skeleton is a clique. (This generalizes ``$G$ chordal $\\Longleftrightarrow$ $G$ splits as a union of chordal graphs that intersect in a common clique''). (3) $Δ$ skeleton-E-chordal $\\Longleftrightarrow$ every nonempty induced subcomplex of $Δ$ has a skeleton-E-simplicial vertex. (Generalizes ``$G$ chordal $\\Leftrightarrow$ every nonempty induced subgraph has a simplicial vertex''.) (4) $Δ$ underclosed $\\Rightarrow$ $Δ$ skeleton-weakly-chordal and weakly-closed. (Generalizes ``$G$ interval $\\Rightarrow$ $G$ chordal and co-comparability''.) (5) All pure E-chordal complexes are vertex-chordal; all pure mid-chordal complexes are weakly-vertex-chordal; all pure very-weakly-chordal complexes are weakly-ridge-chordal. (This expands Bigdeli, Yazdan-Pour and Zaare-Nahandi's work on ridge-chordality.)", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_06ae06b51d", "k": "Computations of $\\tilde{A}_2$ Bruhat intervals via shadows", "domain": "science", "knowledge": "We develop an explicit geometric and algebraic description of shadows in the affine Coxeter complex of type $\\tilde{A}_2$, by introducing a coordinate system based on a decomposition of the complex into tunnels and channels. We also provide an algorithm for converting arbitrary reduced words into coordinates. Using this framework, we identify geometric symmetries of shadows and show that they are governed by the underlying channel structure. This allows us to derive explicit, piecewise formulas for the cardinality of shadows in $\\tilde{A}_2$, depending on the parity of the coordinates. Furthermore, we establish a simple criterion for shadow membership via a counting function that detects admissible positions within channels. These results provide a concrete and computationally effective description of shadows in affine type $\\tilde{A}_2$, bridging the gap between combinatorial definitions and geometric realisations.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_b63c7c3671", "k": "An Isodiametric Theorem and Lattice Diameter-Perfect Codes in $A_3$", "domain": "science", "knowledge": "The root lattice $A_n$, equipped with its graph distance (equivalently, one half of the ambient $\\ell_1$ metric), is isometric to $\\mathbb{Z}^n$ with the asymmetric Manhattan metric. We study two extremal problems in this space -- the isodiametric problem, i.e., determining the maximum anticode cardinality, and the (non)existence of linear diameter-perfect codes, i.e., lattice tilings by optimal anticodes -- and solve them in dimension $3$. We show that, for every integer $D\\ge 0$, the largest cardinality of a diameter-$D$ subset of $A_3$ is $\\binom{D+3}{3}+(D+1)\\lfloor D^2/4\\rfloor$, and this value is attained by the balanced difference of two discrete simplices. We then prove an integrality-refined simplex-packing obstruction: a sublattice of $\\mathbb{Z}^n$ of asymmetric Manhattan distance greater than $D$ induces a lattice packing by $(D+1)Δ_n$ in $\\mathbb{R}^n$. Combining this observation with the exact lattice-packing density of the tetrahedron yields a complete classification in dimension $3$: lattice diameter-perfect codes in $A_3$ exist precisely for $D=1$ and $D=2$. We also give the equivalent statement for perfect $B_h$ sets of cardinality four. Finally, we formulate a conjecture regarding optimal anticodes in arbitrary dimension, and restate it as an intersection problem for uniform multisets.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_0cf36e1331", "k": "Lower bounds on the strength of the determinant", "domain": "science", "knowledge": "We establish new lower bounds for the strength and partition rank of the determinant. For every prime $p$, we prove the exact identity \\[ \\operatorname{str}(\\mathrm{det}_p)=p. \\] A weak monotonicity argument, combined with a bound for gaps between consecutive primes, then gives $\\operatorname{str}(\\mathrm{det}_n)\\ge (1-o(1))n^{0.475}$ for sufficiently large $n$. Since the Birch rank of $\\mathrm{det}_n$ is always $4$, this gives the first explicit family showing that the dependence on the degree in bounds for strength in terms of Birch rank is unavoidable. Viewing $\\mathrm{det}_n$ as an $n$-linear form in its columns, we also prove that its partition rank is at least the largest prime not exceeding $n$. Consequently, \\[ n-n^{0.525}\\le \\operatorname{prk}(\\mathrm{det}_n)\\le n \\] for all sufficiently large $n$, and hence the partition rank of the determinant is $n-o(n)$. The proof introduces an intersection-theoretic method for lower-bounding strength: a short strength decomposition produces a nowhere-vanishing section of a split vector bundle on the complement of the determinantal hypersurface, while a nonzero top Chern class in the Chow ring of $\\mathrm{PGL}_n$ obstructs such a section.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_c9c92f7666", "k": "The optimal constant for minimum weight feedback arc sets in oriented graphs", "domain": "science", "knowledge": "Let $D$ be an oriented graph (a digraph with no directed 2-cycles) with maximum degree $Δ\\ge 1$, equipped with nonnegative arc weights of total weight $w(D)$, and let $\\mathrm{fas}_w(D)$ denote the minimum weight of a feedback arc set of $D$. Alon (2002) proved $\\mathrm{fas}_w(D)\\le(\\frac{1}{2}-\\frac{1}{16\\sqrt{2Δ}})w(D)$. We determine the optimal constant: \\[\\mathrm{fas}_w(D)\\le(\\frac{1}{2}-\\frac{\\sqrt{2}}{6\\sqrtΔ})w(D).\\] In fact, we show a stronger result: $\\mathrm{fas}_w(D)\\le\\frac{1}{2}w(D)-\\frac{\\sqrt{2}}{12}\\sum_v w_2(v)$, where $w_2(v)$ is the $\\ell_2$-norm of the weights of the arcs incident with $v$. Both bounds are attained by the unit-weight directed triangle, so the constant $\\sqrt{2}/6$ is best possible (already among unweighted oriented graphs). The proof combines the vertex-peeling scheme of Berger and Shor with a continuous random-ordering analysis: realizing the random order by independent uniform labels renders the expected local imbalance at each vertex exactly an integrated Khintchine-type functional, and the theorem reduces to the sharp evaluation \\[\\inf_{\\|a\\|_2=1}\\int_0^1 \\mathbb{E}|\\sum_j a_j B_j(q)|\\,dq = \\frac{\\sqrt{2}}{6},\\] where the $B_j(q)$ are i.i.d. Bernoulli$(q)$ random variables, which we prove via Fourier analysis. The proof also yields a randomized, near-linear-time algorithm attaining the bounds in expectation.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_61514d1120", "k": "New excluded minors for the class $\\mathcal{M}_3$ of regular matroids", "domain": "science", "knowledge": "Engel, de Gaay Fortman, and Schreieder attach to each prime $\\ell$ a minor-closed class $\\mathcal{M}_\\ell$ of regular matroids, whose excluded minors govern the failure of the integral Hodge conjecture for curve classes on very general principally polarized abelian varieties. For $\\ell=2$ the class is the cographic matroids, with excluded minors $M(K_5)$ and $M(K_{3,3})$ by Tutte's theorem; for $\\ell=3$ the only excluded minor explicitly identified so far is $M(K_{3,5})$, and the general characterization is [EGFS, Problem 8.8]. We exhibit five new excluded minors for $\\mathcal{M}_3$, of ranks 8, 9, 9, 9, and 10; none contains $M(K_{3,5})$ or any of the other four as a minor, so the excluded-minor list of Problem 8.8 has at least six members, with excluded minors at every rank from 7 through 10. Combined with $M(K_7) \\notin \\mathcal{M}_3$, established by the same authors, and the minor-closedness of $\\mathcal{M}_3$, some minor of $M(K_7)$ of rank between 4 and 6 is a further excluded minor, so the list has at least seven members, six of them explicitly identified. The list is structurally diverse: two bipartite apex constructions with automorphism group $S_4$, and three non-bipartite, non-planar rank-9 graphs with no such apex structure. Every verdict is certified by a finite $\\mathbb{F}_3$-linear-algebra computation with explicit machine-checkable witnesses, cross-checked by independent implementations; the certificates further show that the five matroids lie outside the larger class $\\widetilde{\\mathcal{M}}_3$. We also record what are, to our knowledge, the first $\\ell=5$ data on this corank-8 slice: $M(K_{3,5})$ and the rank-8 and rank-10 minors lie in $\\mathcal{M}_5$ by explicit certificates, and the radical distance $d(M(K_{3,5}))=6$ is determined exactly.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_129a47d7b7", "k": "Closed formulas for energy functions on tensor squares of higher-level perfect crystals in classical affine types", "domain": "science", "knowledge": "For every $l\\geq1$, we construct explicit closed-form coordinate formulas for the local energy functions on the tensor products $B_l\\otimes B_l$ of level-$l$ perfect crystals in classical affine types. A single finite-level max-linear formula covers all seven types: its two branches coincide in type $A_n^{(1)}$, yielding a cyclic maximum of partial sums, whereas in the remaining six types each branch is the maximum of finitely many explicit piecewise-linear expressions in barred coordinates, with type-dependent boundary data. We verify the defining local-energy recursion directly on the finite crystals and derive equivalent recursive forms, allowing the energy to be evaluated without applying the combinatorial $R$-matrix. Substitution into the KMN path character formula gives explicit positive coordinate path sums for the characters of all level-$l$ integrable highest weight modules. After principal specialization, the path exponent can be rewritten as a weighted sum of a position-independent adjacent-pair statistic; comparison with the specialized Weyl--Kac character formula yields a uniform family of level-$l$ Rogers--Ramanujan-type identities equating these sums with explicit infinite products. For a representative low-rank case at level two in each family, we display the complete adjacent-pair degree matrix and list the resulting identities.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_4addb3af9d", "k": "Degree Majorization and Laplacian Eigenvalue Sums for Simplicial Complexes", "domain": "science", "knowledge": "Let $K$ be an $r$-dimensional simplicial complex. We prove that the spectrum of its $(r - 1)$-dimensional up-Laplacian is majorized by the conjugate degree sequence of its $(r - 1)$-dimensional faces: \\[ {\\mathbfλ}_{r-1}(K) \\preccurlyeq {\\mathbf d}_{r-1}^\\top(K). \\] We also establish a Brouwer-type inequality: for every integer $\\ell \\geq 1$, \\[ \\sum_{i = 1}^{\\ell}λ_{r-1,i}(K) \\leq \\frac{r + 1}{2}f_r(K) + \\frac{f_{r - 2}(K)}{r} \\binom{\\ell + 1}{2}, \\] where $λ_{r-1,i}(K)$ denotes the $i$-th largest eigenvalue in the spectrum ${\\mathbfλ}_{r-1}(K)$, and $f_t(K)$ denotes the number of $t$-dimensional faces of $K$. These results provide higher-dimensional analogs of the Grone-Merris-Bai theorem and the Brouwer-Kothari-Tudose theorem and recover the corresponding graph results when $r=1$. We show that the Duval-Reiner conjecture on the majorization by the conjugate degree sequence of vertices fails in every dimension $r \\geq 2$. More precisely, for every $n \\geq r + 5$, we construct a pure $r$-dimensional complex on $n$ vertices that violates the conjectured inequality at the fifth partial sum.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_622316a27c", "k": "Generalized quadrangles with a point-primitive and line-primitive automorphism group with socle $\\PSp_4(q)$", "domain": "science", "knowledge": "Let $\\SA$ be a finite thick generalized quadrangle, and let $G\\leq \\Aut(\\SA)$ act primitively on both points and lines. Building on the almost simple reduction for point-primitive and line-primitive actions, we study the case where the socle of $G$ is the projective symplectic group $\\PSp_4(q)$ with $q\\ge 3$. We show that this hypothesis forces $\\SA$ to be the classical symplectic quadrangle $W(3,q)$ up to duality.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_1d1a745b7d", "k": "Constructions for supersaturation of eventown problems", "domain": "science", "knowledge": "In this paper, we study the supersaturation problems of eventown. Given a family $\\mathcal{A}$ of subsets of an $n$ element set, let op$(\\mathcal{A})$ denote the number of distinct pairs $A,B\\in \\mathcal{A}$ for which $|A\\cap B|$ is odd. We give extremal eventown constructions and show that for fixed $s\\le2^{\\lfloor \\frac{n}{2} \\rfloor}-2$, there exists a collection of $2^{\\lfloor\\frac{n}{2}\\rfloor}+s$ even-sized subsets of an $n$ element set that contains exactly $s\\cdot 2^{\\lfloor \\frac{n}{2} \\rfloor-1}$ pairwise intersections of odd size. This extends the range of $s$ in a conjecture proposed by O'Neill from $2^{\\lfloor \\frac{n}{2} \\rfloor}-2^{\\lfloor \\frac{n}{4} \\rfloor}$ to $2^{\\lfloor \\frac{n}{2} \\rfloor}-2$. We also give a construction using symmetric designs to prove that when $k$ is even and $4k-1$ is a prime power, there exists a collection of $2^{\\lfloor\\frac{4k-1}{2}\\rfloor}+s$ even-sized subsets of a $4k-1$ element set $\\mathcal{A}_s$ with $op(\\mathcal{A}_s)=s \\cdot 2^{{\\lfloor\\frac{4k-1}{2}\\rfloor}-1}$, $1\\leq s\\leq4k-1$.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_16bedf37bc", "k": "Flag Space, Matroidal Schur Algebras and the Steinberg Representation", "domain": "science", "knowledge": "We define via generators and relations an extended version of the matroidal Schur algebras introduced in earlier work of the author with Tom Braden. In the case of projective geometries, we relate the representation theory of the resulting algebras to the structure of the modular Steinberg representation of $GL(n,q)$. In the general case we show that the extended matroidal Schur algebras have a triangular, or Reedy, decomposition and are thus quasi-hereditary. We also obtain a new description of the original matroidal Schur algebras and clarify the connection to the flag spaces of Brylawski--Varchenko.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_92a4bc7ea8", "k": "Multiplier obstructions for Legendre pairs of length 333", "domain": "science", "knowledge": "A Legendre pair of length 333 would yield a Hadamard matrix of order 668, the smallest order presently unresolved by the Hadamard conjecture. We study the structured case in which both sequences are fixed by a common subgroup $H\\leq(\\mathbb Z/333\\mathbb Z)^\\times$ acting by coordinate multiplication. We prove that such a pair can exist only when $|H|\\leq 6$. After a mod-3 compression reduces the problem to an order-108 kernel, there are exactly 30 subgroups. We exclude 21 of them, including all 19 subgroups of order at least 9. The final order-9 subgroup is eliminated analytically: its orbit structure restricts the 9-compressed entries to $\\{\\pm1,\\pm17,\\pm19,\\pm35,\\pm37\\}$; the Legendre equations force a $+17,-17$ pair in one compressed sequence, and a single-shift autocorrelation bound then contradicts the required compressed correlation. The remaining exclusions use full-image compression, a row-sum congruence, exact meet-in-the-middle enumeration, and proof-carrying pseudo-Boolean encodings. The solver-assisted cases are accompanied by independently checked DRAT proofs or direct arithmetic certificates. The result constrains fixed common-multiplier symmetry only; the unrestricted existence problems remain open.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_ae2ff9ffc1", "k": "Infinite Sidon-type sets for zero-sum linear forms", "domain": "science", "knowledge": "Let $h \\geq 2$, and let $\\mathbf{b} = (b_1,\\dots,b_h)\\in \\mathbb{Z}^h$ be a zero-sum vector with nonzero coordinates. For a set $A=\\{a_1<a_2<\\cdots\\}\\subseteq\\mathbb{N}$, let $r_{A,\\mathbf{b}}(n)$ denote the number of $h$-tuples $(x_1,\\ldots,x_h)$ of pairwise distinct elements of $A$ satisfying $b_1x_1+\\cdots+b_hx_h=n$. We study density restrictions on sets $A$ for which these representation counts remain small, obtaining analogues of the classical density theorem for infinite Sidon sets. In the case $\\mathbf{b} = (c_1,-c_1,\\dots,c_k,-c_k)$, we prove that if $A(x)/(x/\\log x)^{1/2k}\\to\\infty$, then $\\frac{1}{x}\\sum_{|n|\\leq x} r_{A,\\mathbf{b}}(n)\\to\\infty$, whereas if $A(x)\\gg x^{1/2k}$, then $\\frac{1}{x}\\sum_{|n|\\leq x} r_{A,\\mathbf{b}}(n)\\gg\\log x$. This recovers Chen's theorem on $B_{2k}$-sequences. For general zero-sum vectors $\\mathbf{b}$, we prove analogous bounds under gap conditions: if $a_{n+1}-a_n=o(n^{h-1}\\log n)$, then $\\frac{1}{x}\\sum_{|n|\\leq x} r_{A,\\mathbf{b}}(n)\\to\\infty$, whereas if $a_{n+1}-a_n\\ll n^{h-1}$, then $\\frac{1}{x}\\sum_{|n|\\leq x} r_{A,\\mathbf{b}}(n)\\gg\\log x$.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_55950bb332", "k": "An Improved Upper Bound for Colorings Without Symmetrically Colored $k$-Term Arithmetic Progressions", "domain": "science", "knowledge": "Given a coloring $c$ and an even $k\\ge 4$, a nontrivial $k$-term arithmetic progression~($k$-AP) $a,a+d,\\ldots,a+(k-1)d$ is called symmetrically colored if $c(a+(i-1)d)=c(a+(k-i)d)$, $\\forall i\\in[k/2]$. Deng, Tidor, and Zhao asked whether $[N]$ admits a coloring with $N^{o(1)}$ colors and no such 4-APs, and gave an $O(N^{\\log_{22}3})$-coloring of $[N]$. We give an $O_k(p)$-coloring of $\\mathbb Z/p^{k^2/4}\\mathbb Z$ without such $k$-APs for every even $k\\ge 4$ and every prime $p>k$, and hence an $O(N^{4/k^2})$-coloring of $[N]$, improving the exponent in the upper bound for $4$-APs from $\\log_{22}3$ to $1/4$. The construction combines a carry-control coloring of base-$p$ digits with a layered field norm mapping. Together with Behrend-style product colorings, our result for $4$-APs gives $h(N)\\leq N^{1/4+o(1)}$ in Erdős's Problem~160 on coloring every nontrivial 4-AP with at least three colors. This result also yields $ρ_4(α)=O_\\varepsilon(α^{5-\\varepsilon})$ for every $\\varepsilon>0$, improving the bound toward Ruzsa's question. Our result for $k$-APs disproves Gowers' conjectured lower bound for all even $k\\ge6$ for the first time.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_9f98d8becc", "k": "Pattern avoidance in permutations and their rotations", "domain": "science", "knowledge": "A rotation of a permutation is a new permutation obtained by moving the first several terms of the permutation to the end of the permutation. A circular permutation is the set of all rotations of a permutation. The enumerations of permutations and circular permutations avoiding patterns of length three and four are well studied. In this paper, we provide exact formulas for the number of permutations whose first $k\\geq 2$ rotations all avoid a given pattern of length three, as well as the number of permutations whose first three rotations respectively avoid the rotations of a given pattern of length three. In contrast to permutations and circular permutations avoiding patterns of length three, the Wilf-equivalence classes under study are entirely determined by complements and reverses. We also classify and enumerate permutations whose first two rotations avoid different patterns of length three.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_982538345b", "k": "Application of the Combinatorial Nullstellensatz to magic-type graph labelings", "domain": "science", "knowledge": "Let $G=(V,E)$ be a simple graph, and let $k\\geq 2$ be an integer. For an edge labeling $h:E(G)\\to \\mathbb{Z}_{k} \\backslash \\{0\\}$, define the induced vertex label by \\[ h^+(v)=\\sum_{e \\ni v} h(e) \\pmod{k}. \\] For $t\\in \\mathbb Z_k$, we say that $G$ is \\emph{$t$-sum $\\mathbb Z_k$-magic} if there exists such a labeling $h$ satisfying \\[ h^+(v)=t \\qquad\\text{for all }v\\in V. \\] We say that $G$ is \\emph{$\\mathbb Z_k$-magic} if $G$ is $t$-sum $\\mathbb Z_k$-magic for some $t\\in \\mathbb Z_k$. Similarly, if there exists an edge labeling $h: E(G) \\to \\mathbb{Z}_{k} \\backslash \\{0\\}$ such that the induced vertex labeling $h^+(v)=\\sum_{e\\ni v} h(e)$ (mod $k$) is injective, then $G$ is called \\emph{$\\mathbb{Z}_{k}$-antimagic}. In this paper, we use the Combinatorial Nullstellensatz to analyze these two types of magic graph labelings.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_996b5169fb", "k": "Shellability of relative squeezed balls and spheres", "domain": "science", "knowledge": "Squeezed balls and spheres, introduced by Kalai, form a rich class of triangulated complexes arising from subcomplexes of cyclic polytopes, with well-understood shellability properties. Recently, Novik and Zheng introduced relative squeezed balls, obtained as differences of squeezed balls, and used them to construct large families of highly neighborly simplicial spheres. While these complexes are known to be constructible, their shellability has remained open. In this paper, we resolve this question by proving that both relative squeezed balls and their boundary complexes are shellable. We provide explicit shelling orders and characterize restriction faces, thereby establishing strong combinatorial structure for this new class of complexes.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_ccd15ab58b", "k": "Combinatorial formula for the Moore-Penrose inverse of the complex signless Laplacian of an oriented graph", "domain": "science", "knowledge": "We find necessary and sufficient conditions for the rank of the signless incidence matrix of a weakly connected oriented graph with non-zero complex edge weights. We use this to find the combinatorial formulas for the Moore-Penrose inverse of the complex signless incidence and complex signless Laplacian matrix of a weakly connected oriented graph with non-zero complex edge weights. This resolves the open problem posed in the concluding remarks of Barik et. al. (Discrete Mathematics 349 (9), 115117, 2026).", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_e3c5a6cc42", "k": "Comments on \"Condensation of the digraph associated with a reciprocal matrix and a vector\", arXiv:2607.10279", "domain": "science", "knowledge": "Since manuscript \\cite{R}, recently posted on arXiv, is closely related to our previous work, we would like to clarify some facts about it. Namely, we would like to point out that several results in \\cite{R} are already known (Theorem 1 and Corollaries 2, 3 and 4) and some observations regarding our work \\cite{FJ} are not correct. We also notice that Rédei's Theorem, a classical result in graph theory, gives a simple tool to show Theorems 5 and 9.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_32364559ba", "k": "The sharp exponent for the minimal distance problem", "domain": "science", "knowledge": "We show that for every fixed $\\varepsilon>0$, there exist arbitrarily large families of point-line pairs $(x_1,\\ell_1),\\ldots,(x_n,\\ell_n)$ in $[0,1]^2$, with $x_i \\in \\ell_i$ for all $i$, and such that $\\operatorname{dist}(x_i,\\ell_j)\\ge n^{-2/3-\\varepsilon}$ for all $i \\neq j$. Combined with a previous result of Cohen, the author, and Zakharov, this solves the minimal distance problem.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_f90cbbddee", "k": "A 13-vertex counterexample to $e$-log-concavity for chromatic quasisymmetric functions", "domain": "science", "knowledge": "We exhibit a connected natural unit interval graph on $13$ vertices whose chromatic quasisymmetric function has an elementary-basis coefficient that is positive, palindromic, and unimodal, but not log-concave. For the Hessenberg function $h=(2,4,4,6,7,10,10,10,10,12,12,13,13)$ and $λ=(6,5,1,1)$, the coefficients of $q^5,q^6,q^7$ in $[e_λ]X_{G(h)}(\\mathbf{x};q)$ are $1,6,38$; hence $6^2<1\\cdot38$. This disproves Conjecture 5.3 of Abreu--Nigro and its later formulations by Tom and by Sagan--Tom. The calculation is certified by a standalone deterministic verifier using exact integer and rational arithmetic.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_8946005b11", "k": "The inverse problem for the Steiner-Wiener index of trees", "domain": "science", "knowledge": "For a connected graph $G$ and a set $S\\subset V(G)$, the Steiner distance $d_G(S)$ is the minimum number of edges in a connected subgraph of $G$ containing $S$. The Steiner-Wiener $k$ index is defined by $\\mathrm{SW}_k(G) = \\sum_{S\\subset V(G), |S|=k} d_G(S)$. We study the inverse problem for this invariant restricted to trees: for fixed $k$, which positive integers occur as $\\mathrm{SW}_k(T)$ for a finite tree $T$? We prove that all sufficiently large positive integers occur as $\\mathrm{SW}_k(T)$ for some finite tree $T$ if and only if $k$ is even. For odd $k$, we further show that the set of attainable values has asymptotic density of order $k^{-δ}(\\log k)^{-3/2}$, where $δ$ is the Erdős-Tenenbaum-Ford constant.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_34302c0402", "k": "The Erdős-Lovász Tihany Conjecture holds for all even-hole-free graphs", "domain": "science", "knowledge": "Let $s, t\\ge2 $ be integers. A graph $G$ is $(s,t)$-splittable if $V(G)$ can be partitioned into two sets $S$ and $T$ such that $χ(G[S ]) \\ge s$ and $χ(G[T ]) \\ge t$. The Erdős-Lovász Tihany Conjecture from 1968 asserts that every graph $G$ satisfying $ω(G)<χ(G)=s+t-1$ is $(s,t)$-splittable. A vertex of a graph is bisimplicial if the set of its neighbors can be expressed as the union of two cliques. Let $G$ be a graph with $ω(G)<χ(G)=s+t-1$. We prove that if $G$ does not contain $C_4$ as an induced subgraph and every induced subgraph of $G$ has a bisimplicial vertex, then $G$ is $(s,t)$-splittable. Combining our result with a recent result of Chudnovsky and Seymour, which states that every non-empty even-hole-free graph has a bisimplicial vertex, we obtain that the Erdős-Lovász Tihany Conjecture holds for all even-hole-free graphs.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_a9b2a660d5", "k": "On (3,1)-regular graphs with one more vertex than edges", "domain": "science", "knowledge": "Sequence A339987 of the OEIS counts (3,1)-regular graphs having one more vertex than edges by half the number of vertices. A recurrence relation satisfied by this sequence was guessed by Kauers and Koutschan in 2023. We confirm it in three ways: first, by a representation as the diagonal of a triple sum and an elaborate variant of traditional creative telescoping that makes an a posteriori validation possible; second, by a residue representation and a direct calculation by reduction-based creative telescoping; third, by a combinatorial recurrence on graph families and a calculation by differential elimination. Each of those three approaches leads to a formally complete proof and involves a computer calculation in one way or another.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_ebaa3709c6", "k": "Triple arrays from ovals in finite projective planes", "domain": "science", "knowledge": "In this paper we prove that whenever a projective plane of odd order n contains an oval, it may be used to construct a triple array with n + 1 rows, n.n columns and n(n + 1) symbols. In particular, for any odd prime power s, we use the projective plane over the field of order s to construct an explicitly given (s + 1) times s.s triple array on s(s + 1) symbols. This is only the third infinite series of triple arrays (up to transposition) to be found. Our construction proves a recent conjecture of Gordeev and Ohman (2025) in the case of odd prime powers.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_ef122feb83", "k": "Generic properties of discrete Steklov eigenfunctions", "domain": "science", "knowledge": "Let $G=(V,E)$ be a finite connected graph with boundary $B$. We prove that for a generic positive edge weight function $w \\in \\mathbb{R}^{|E|}$, the Steklov eigenvalues of $(G,B,w)$ are simple and every Steklov eigenfunction does not vanish on the boundary. More precisely, the exceptional weights are contained in a zero set of a non-identically zero polynomial and hence form a set of Lebesgue measure zero and Hausdorff dimension at most $|E|-1$. Our results provide a discrete extension of the genericity theorem for the Steklov problem on compact manifolds.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_33c8f83f16", "k": "A Matrix-Degree Obstruction to Rational Generation of Boolean-Lattice Pseudo-Roots", "domain": "science", "knowledge": "For the neighborhood seed associated with the four-vertex path $P_4$, the diamond operations do not recover all Boolean-lattice pseudo-roots. The corresponding question for unrestricted rational operations in the free skew field is subtler: the seed map has an invertible linearization and therefore a unique formal inverse near every generic scalar point. We prove that this formal inverse is not free rational. A symmetric one-parameter curve of $2 \\times 2$ matrix outputs has a formal inverse whose coefficient field contains an element of degree three over $\\mathbb{Q}(t)$. An exact elimination in a quadratic Pauli algebra produces the irreducible cubic. Its conjugate inverse branches are unramified, forcing the generic matrix degree of the seed map to be at least three in every size $n \\ge 2$. This contradicts the degree-one consequence of any free rational inverse. The same matrix-degree argument, without specializing a hypothetical inverse, extends the obstruction to every graph containing an induced $P_4$.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_e533e88fa9", "k": "Graphon as a Bridge between Graphs and Manifolds", "domain": "science", "knowledge": "We show that there exist graphons that interpolate between Riemannian manifolds and weighted geometric graphs. Specifically, the graph-to-manifold approximation used in manifold learning can be regarded as the composition of a graph-to-graphon convergence and a graphon-to-manifold convergence in a certain sense. Furthermore, we establish a monotonicity inequality which reveals an implicit relationship between numerous combinatorial parameters and geometric quantities on graphons. Using this inequality, we find relations among conductance, maxcut problem, capacity, and packing radius, as well as their limiting behaviours under graph-to-graphon and graphon-to-manifold convergences; some of these relations are novel even for simple graphs and closed manifolds.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_874523d2ed", "k": "Matchings and Near-Optimal 2-Factor Packings in Percolated Vertex-Transitive Graphs", "domain": "science", "knowledge": "Let $G$ be a connected simple vertex-transitive graph on $n$ vertices with degree $d$, and let $G_p$ be the random spanning subgraph obtained by retaining each edge independently with probability $p$. Motivated by a conjecture of Bedert, Draganić, Müyesser, and Pavez-Signé on Hamilton cycles in percolated Cayley graphs, we establish its matching and $2$-factor consequences for the larger class of all connected vertex-transitive host graphs. First, for every $A>0$, there is $C=C(A)>0$ such that, for every $N\\ge n$, the condition $(1-p)^d\\le N^{-C}$ implies, with probability at least $1-N^{-A}$, that $G_p$ has a perfect matching when $n$ is even and that $G_p-v$ has a perfect matching for every vertex $v$ when $n$ is odd. Second, if $ν_2(H)$ is the maximum number of pairwise edge-disjoint spanning $2$-factors in $H$, then, for every $A>0$ and $0<ε<1$, \\[ ε^2pd\\ge64(A+6)\\log(2n) \\] implies \\[ \\mathbb P\\left( ν_2(G_p)\\ge\\left\\lfloor(1-ε)\\frac{pd}{2}\\right\\rfloor \\right)\\ge1-n^{-A}. \\] Consequently, $pd\\ge C\\log n$ guarantees both the appropriate matching property and a spanning $2$-factor with high probability, uniformly over all connected vertex-transitive graphs. If $pd/\\log n\\to\\infty$, then $ν_2(G_p)=(1+o(1))pd/2$ with high probability. The coefficient $1/2$ is best possible because each spanning $2$-factor contains $n$ edges, whereas $G_p$ contains about $pnd/2$ edges.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_e71435f598", "k": "Induced Subgraph Bounds on the Zero Forcing Number and a $(χ, ω, Z)$-Conjecture", "domain": "science", "knowledge": "Let $G$ be a graph with chromatic number $χ(G)$, clique number $ω(G)$ and zero forcing number $Z(G)$. We establish new lower bounds on $Z(G)$ in terms of induced triangle-free subgraphs. In particular, we show that if a graph $G$ contains an induced triangle-free subgraph $H$ with minimum degree $δ(H) \\ge 3$, then $Z(G)\\geδ(H)+1$. Motivated by this result and the bound $χ(G) \\leq Z(G) + 1 $ by Taklimi (2013), we conjecture that \\begin{equation*} χ(G) \\leq \\left \\lceil \\frac{ω(G)+Z(G)+1}{2}\\right\\rceil. \\end{equation*} As supporting evidence, we prove that the conjecture holds for triangle-free regular graphs and also provide numerical evidence.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_524066aeff", "k": "Generalised flip order on the faces of nestohedra", "domain": "science", "knowledge": "Classical shuffle products on permutations and binary planar rooted trees (i.e., on the vertices of permutohedra and associahedra) admit descriptions in terms of intervals in the weak Bruhat order and the Tamari order, respectively. Palacios and Ronco extended these products as well as their interval description to surjections and planar rooted trees (i.e. on all faces of permutohedra and associahedra). In this article, we present a broad generalisation of this phenomenon. We show that the shuffle product on faces of certain families of nestohedra admits an interval description with respect to the generalised flip order, a partial order defined on the faces of nestohedra through elementary splitting and fusion operations on the tree-like combinatorial objects encoding them. The generalised flip order extends the flip order of Barnard and McConville from vertices to all faces of nesthedra. We further compare it with the facial weak order of Dermenjian-Hohlweg-Pilaud and the generalised Tamari order of Ronco, and we provide its characterisation in terms of (generalised) inversions.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_fab7d81743", "k": "Combinatorial geometry of the 2D Toda lattice and Davey Stewartson equation", "domain": "science", "knowledge": "The KP equation is a prototypical $(2+1)$-dimensional integrable PDE. Its soliton solutions are famously parametrized by the Sato Grassmannian. In seminal work, Kodama and Williams made the surprising discovery that the combinatorics of soliton solutions are intimately related to the combinatorics of the totally positive Grassmannian as pioneered by Postnikov. They introduced novel algorithmic methods inspired by polyhedral structures arising from tropical geometry. Soliton solutions to the 2D Toda lattice and the Davey--Stewartson equation, two closely related integrable systems with soliton solutions, are also classified by the Sato Grassmannian. Kodama suggested that the methods of his work with Williams could generalize to these two integrable equations. In this work, we show that this is indeed the case. We derive algorithms to produce contour plots from elements in the totally nonnegative Grassmannian in both cases. In the asymptotic setting, we recover and refine previous work of Biondini and Wang; as well as Biondini, Kireyev and Maruno.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_5d4ecaaf17", "k": "The Duval--Reiner Conjecture: Counterexamples and the Second Partial-Sum Inequality", "domain": "science", "knowledge": "Let \\(F\\subseteq\\binom{V}{q}\\) be a \\(q\\)-uniform family on a finite vertex set \\(V\\). Write \\(s_r(F)\\) for the sum of the \\(r\\) largest eigenvalues of its simplicial up-Laplacian and \\(d_F(v)\\) for the degree of \\(v\\in V\\). Then $D_r(F)=\\sum_{v\\in V}\\min\\{d_F(v),r\\}$ is the \\(r\\)-th partial sum of the conjugate degree sequence of \\(F\\). The majorization assertion in the Duval--Reiner conjecture [Trans. Amer. Math. Soc., 2002] states that \\(s_r(F)\\le D_r(F)\\) for every \\(q\\)-uniform family \\(F\\) and every \\(r\\ge1\\). We disprove this assertion in two complementary senses: for every \\(r\\ge5\\), there is a strict counterexample at index \\(r\\) in some uniformity, while every uniformity \\(q\\ge3\\) admits a strict counterexample at some index \\(r\\ge5\\). In contrast, we prove the universal inequality \\(s_2(F)\\le D_2(F)\\) and classify all equality cases. The counterexamples are obtained from two \\(3\\)-uniform seeds with explicitly computed characteristic polynomials through defect-preserving ridge-whiskering and set-complement duality. For the second partial sum, core completion reduces the problem to the boundary matrix of a complete simplex, where Ky Fan variational and compression arguments yield both the inequality and its equality classification.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_5a30465a1e", "k": "Restricted $r$-Stirling numbers of the second kind", "domain": "science", "knowledge": "In this paper, we introduce the $(r,d)$-Stirling numbers of the second kind, a new generalization of the $r$-Stirling numbers obtained by imposing a distance restriction: for any two elements $i$ and $j$ within the same subset, we require that $|i-j|>d$. We establish recurrence relations, an explicit formula, ordinary and exponential generating functions, a reduction identity connecting these numbers to the standard $r$-Stirling numbers, and several combinatorial identities.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_5fc97d63b1", "k": "How Close is a Tree to a Euclidean Minimum Spanning Tree?", "domain": "science", "knowledge": "Let $Γ$ be a straight-line crossing-free drawing of a tree $T$. A \\emph{bad pair} in $Γ$ is a pair of non-adjacent vertices of $T$ whose Euclidean distance in $Γ$ is smaller than the length of the longest edge in the path connecting them in~$Γ$. When $Γ$ has no bad pairs, $Γ$ is a Euclidean Minimum Spanning Tree of its vertex set (or EMST-drawing for short). Deciding whether a tree of maximum degree at most six admits an EMST-drawing is known to be \\NP-hard. In contrast, we characterize those caterpillars that admit an EMST-drawing. The characterization gives rise to a linear-time algorithm that decides if a caterpillar admits an EMST-drawing, and in the affirmative case, computes such a drawing. For caterpillars of maximum degree six, we further present a linear-time algorithm to compute a crossing-free straight-line drawing with the minimum number of bad pairs. For $n$-vertex trees with maximum vertex degree $Δ$, we prove the $Δ^2n\\log n$ upper bound on the minimum number of bad pairs. In the special case of stars, we construct a drawing with the minimum number of bad pairs.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_fc80542eaf", "k": "Exact certification of the coordinate fields of the triangle-free Exoo-Ismailescu unit-distance graphs EI17 and EI19 (HoG 51375, 51376): a solvable-non-solvable dichotomy (origami vs. S20) and the Laman-number conjecture", "domain": "science", "knowledge": "We certify, exactly, the coordinate fields of a faithful planar realization of two neighbouring triangle-free Exoo-Ismailescu unit-distance graphs (UDGs), and show they realize the two opposite extremes of the constructibility hierarchy. The 17-vertex graph EI17 (House of Graphs 51375) is the smallest triangle-free UDG with chromatic number 4; the 19-vertex graph EI19 (HoG 51376) is its state-of-the-art origami neighbour. In both, fixing a rational base edge, the remaining vertices are intersections of unit circles -- each on the radical axis of its two neighbours, a tower of square roots over the free angles -- and a small closure system locks the realization. For EI19 the base lies in Q(sqrt 2, sqrt 5, sqrt 7) and a single free angle has an irreducible degree 12 = 2^2*3 minimal polynomial with Galois group the solvable transitive group 12T236 (order 2304 = 2^8*3^2): not ruler-and-compass, but origami-constructible (the cubic Beloch fold O6 necessary, in casus irreducibilis). For EI17 two free angles are locked by two closures, whose resultant is irreducible of degree 20 = 2^2*5 with Galois group the full symmetric group S20 (a Frobenius census exhibits a 17-cycle, forcing A20 by Jordan, and an odd 20-cycle, raising it to S20): non-solvable, so the coordinates are not expressible by radicals -- neither compass nor origami of any fold order. Thus the smallest triangle-free 4-chromatic UDG is the generic, maximally exotic case, the exact opposite of its origami neighbour. We give the full certification pipeline as explicit algorithms, record two methodological pitfalls, and read the pair through a conjectural bridge between the Laman number and the Galois group.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_cd5c3b2a73", "k": "On the possible cardinalities of quantum Latin squares", "domain": "science", "knowledge": "In this paper, we show that for any integer \\(v \\ge 8\\) with \\(v \\ne 9,11,23\\), a quantum Latin square of order $v$ exists for every cardinality \\(c \\in [v,\\,v^{2}] \\setminus \\{\\,v+1\\,\\}\\).", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_5832630c59", "k": "The Fibonacci Rectangle Game: Two First-Move Classes and a Triangle-Induced Choice", "domain": "science", "knowledge": "A square-adjoining rectangle game generates the Fibonacci numbers and the Fibonacci spiral from a simple geometric rule. If one starts from a square, the four possible first moves are all equivalent by rotation. If one starts instead from a non-square rectangle, there are still four geometric placements for the first square, but they split into exactly two equivalence classes: long-side-first and short-side-first. We show that both classes are governed by the same Fibonacci-type recursion with different initial conditions, and that in both cases the successive aspect ratios converge to the golden ratio (phi). We then add a brief geometric remark: the Hypotenuse-Axis Intercept (HAI) construction from a right triangle produces a natural ordered seed whose outward and inward branches determine precisely those two first-move classes.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_4f69e64f59", "k": "The chromatic number of Euclidean space with dense color classes", "domain": "science", "knowledge": "In this note we construct colorings of Euclidean space $\\mathbb{R}^n$ with finitely many colors such that any two points at unit distance have different colors and, in addition, each color class is dense in $\\mathbb{R}^n$. In particular, 12 dense colors suffice to color $\\mathbb{R}^2$. In arbitrary dimension, we show that $nχ(\\mathbb{R}^n)+1$ colors suffice, where $χ(\\mathbb{R}^n)$ denotes the chromatic number of $\\mathbb{R}^n$ in the standard formulation.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_cb84feef9f", "k": "On Boolean sublattices of finite partition lattices", "domain": "science", "knowledge": "We investigate maximal Boolean sublattices of the partition lattice Part(U) of a finite universe U. First, any largest size Boolean sublattice of Part(U) can be formed using all partitions whose blocks are subtrees of a tree with vertex set U. It is shown that a maximal Boolean sublattice of Part(U) always contains the least and the largest elements of Part(U). Boolean sublattices B of PartU containing 0 are characterized by a certain condition imposed on the cycles of a linear hypergraph corresponding to the atoms of B on the set U. We show that B can be extended to a Boolean sublattice of Part(U) with a largest size, if and only if the hypergraph induced by its atoms is a hypertree. This is the case when B is formed by all the partitions whose blocks are intervals in a generalized sense. The main result states that all partition lattices of height at least three have maximal Boolean sublattices for all possible dimensions at least three", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_8b7e869bff", "k": "Higher cluster tilting objects in locally finite triangulated categories", "domain": "science", "knowledge": "We study higher cluster tilting objects through covering functors from derived categories of hereditary algebras. The covering formalism reduces the existence problem to the equivariant problem of finding \\(G\\)-stable \\(d\\)-cluster tilting objects in \\(d\\)-cluster categories. For triangulated categories with finitely many indecomposable objects this gives a complete ADE existence criterion and explicit counting formulas. We also prove that, in finite Frobenius models, the endomorphism algebras of all \\(d\\)-cluster tilting objects are derived equivalent. Applications are given to Cohen--Macaulay finite categories, finite noncommutative crepant resolutions, and rigidity dimensions of representation-finite self-injective algebras.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_2f1ac58a8d", "k": "Hamiltonian cycles in 7-tough $(P_4\\cup P_1)$-free graphs", "domain": "science", "knowledge": "Shan~[J. Graph Theory (2026)] proved that every 23-tough $(P_4\\cup P_1)$-free graph on at least three vertices is Hamiltonian. We improve this bound to 7 by replacing the final cut analysis in Shan's framework with an asymmetric separation criterion and a cograph covering lemma.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_5d60bd2ffe", "k": "Negative Latin-Square-Type Partial Difference Sets in Non-Elementary Abelian 2-Groups", "domain": "science", "knowledge": "Using cubic cyclotomic classes, character theory, and product constructions motivated by generalized Denniston partial difference sets, we construct negative Latin-square-type partial difference sets in $\\mathbb{F}_{2^6}^{+}\\times\\mathbb{Z}_4^4$, $\\mathbb{Z}_4^4\\times\\mathbb{F}_{2^{10}}^{+}$, $\\mathbb{F}_{2^{10}}^{+}\\times\\mathbb{Z}_8^4$, $\\mathbb{Z}_8^4\\times\\mathbb{F}_{2^{14}}^{+}$, and $\\mathbb{Z}_4^4\\times\\mathbb{Z}_{16}^2$. The first four constructions replace cubic cyclotomic partitions by partitions of non-elementary abelian 2-groups having the same character-value patterns. To the best of our knowledge, these are the first partial difference sets with the stated parameters in the indicated groups.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_4416e7f853", "k": "Polyhedral Maps of Cubic Graphs with given Automorphism Groups", "domain": "science", "knowledge": "L. Babai introduced a method for constructing a cubic graph whose automorphism group is isomorphic to a given finite group $G$, obtained by modifying a corresponding Cayley graph of $G$. Building on this approach, we construct a cubic graph that admits a polyhedral map whose automorphism group, as well as the automorphism group of the polyhedral map itself, is isomorphic to $G$.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_2526f3c4f1", "k": "Energy and independence number", "domain": "science", "knowledge": "For a graph $G$ of order $n$, with adjacency eigenvalues $λ_1(G) \\geq \\cdots \\geq λ_n(G)$, the \\emph{energy} of $G$ is defined to be \\[\\mathcal{E}(G)=\\sum_{i=1}^{n} |λ_i(G)|.\\] A well-known conjecture from the 1980s by Fajtlowicz states that for any graph $G$, \\[\\mathcal{E}(G) \\ge 2\\left(n-α(G)\\right),\\] where $α(G)$ denotes the independence number. We prove this conjecture.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_f8a035fbb8", "k": "A direct injection for the strong $q$-log-convexity of Touchard polynomials", "domain": "science", "knowledge": "We provide a direct injection for the well-known strong log-convexity of the Bell numbers $B_n$, that is $B_mB_n\\le B_{m-1}B_{n+1}$ for every $1\\le m\\le n$. Our injection $Π_m\\timesΠ_n\\to Π_{m-1}\\timesΠ_{n+1}$, where $Π_n$ denotes the set of all partitions of $[n]$, preserves the total number of blocks in the pair of partitions. In other words, it is also an injection for the strong $q$-log-convexity of Touchard polynomials, a result established by Chen, Wang, and Yang using analytical arguments. As an application of the injection, we also recover a related result of Chern, Diaconis, Kane, and Rhoades.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_a11bdcd276", "k": "A Sharp Ramsey Theorem for Admissible Colorings of Ordered Cliques", "domain": "science", "knowledge": "Let \\(f(k)\\) be the minimum integer \\(N\\) such that any red--blue edge-coloring of the ordered complete graph on \\(N\\) vertices contains a set of \\(k\\) vertices whose induced coloring is admissible. In this note, we obtain the exact value of $f(k)$ for $k\\ge 3$, which confirms a conjecture posed by Bradač, Liu, Wu and Xu.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_eb22d2c805", "k": "Classification of two-distance-transitive Cayley graphs of the semi-dihedral groups", "domain": "science", "knowledge": "The class of 2-distance-transitive graphs naturally generalizes distance-transitive graphs and plays a central role in algebraic graph theory. Classifying such graphs for a prescribed underlying group is a key open problem. A vertex-transitive graph $Γ$ is said to be $2$-distance-transitive if, for each $i\\in \\{1,2\\}$, any two pairs of vertices with identical distance $i$ in $Γ$ can be mapped to each other via some automorphism of the graph. In this paper, we present a complete classification of all $2$-distance-transitive Cayley graphs of the semi-dihedral groups.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_79ee2f4c85", "k": "Hitting time mixing for random $k$-cycles", "domain": "science", "knowledge": "In this paper, we study the random walk on the symmetric group $\\mathfrak{S}_n$ generated by the conjugacy class of $k$-cycles, where $2\\le k=o(n/(\\log n)^4)$. We prove that the walk exhibits hitting-time mixing: at the first time when every card has been touched, the distribution is already close to equilibrium. For odd $k$, the equilibrium measure is the uniform measure on $\\mathfrak{A}_n$. For even $k$, the walk first mixes to the parity mixture determined by the hitting time, and in our range this mixture is asymptotically $U_{\\mathfrak{S}_n}$. Our argument combines a refined fixed-time approximation for the random $k$-cycle walk near the cutoff window with an auxiliary marking scheme inspired by Jain-Sawhney's work (arXiv:2410.23944) on random transpositions. The main new feature is a parity-compatible coupling which handles both odd and even $k$-cycles in a unified framework. We also prove a hitting-time mixing result in the opposite regime $k\\ge n-o(n^{1/2})$, and formulate a conjecture for all $2\\le k\\le n-1$.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_2801d67fd4", "k": "Negative-index $u$-deformed homogeneous functions and Ramanujan partial theta functions", "domain": "science", "knowledge": "We introduce and study a class of \\(u\\)-deformed homogeneous functions of complex order \\(\\rr_α(x,y;u\\,|\\,q)\\), obtained by extending \\(u\\)-deformed homogeneous polynomials beyond the nonnegative integer regime. We establish their basic analytic and algebraic properties, including convergence criteria, recurrence relations, \\(q\\)-derivative formulas, and an operator-theoretic realization through a deformed \\(q\\)-exponential operator. A central feature of the theory appears in the negative-index sector: for every nonnegative integer \\(n\\), the functions \\(\\rr_{\\minus n\\minus1}(x,y;u\\,|\\,q)\\) admit finite triangular decompositions in terms of shifted Ramanujan partial theta functions. We also prove the inverse triangular transformation, showing that the corresponding partial theta functions can be recovered from the negative-index \\(u\\)-deformed homogeneous functions. We also derive pantograph-type \\(q\\)-difference equations for these functions. Finally, we derive specialized forms of the negative-index theory in Cauchy and Stieltjes--Wigert-type regimes, including a boundary geometric case and genuine partial-theta regimes.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_b9b56f5d04", "k": "Rational exponents near 3/2", "domain": "science", "knowledge": "Given a graph $H$, the extremal number $ex(n,H)$ is the maximum number of edges in an $n$-vertex graph not containing $H$ as a subgraph. The well-known rational exponents conjecture of Erdős and Simonovits states that for any rational $γ\\in (1,2)$ there exists a single bipartite graph $H$ satisfying $ex(n,H)=Θ(n^γ)$. Among other results, the conjecture has been verified for all $γ=1+a/b$, where $b>a^2$, by Jiang and Qiu and for all $γ=2-a/b$, where $b>\\max\\{a, (a-1)^2\\}$, by Conlon and Janzer. In this paper, we establish the rational exponents conjecture for many $γ$ near the center of the interval, namely, for all $γ=1+\\frac{rt-1}{2rt+2r}$, where $r,t$ are natural numbers satisfying $t\\geq 2$, $r\\geq 2t+3$.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_e83ba8646c", "k": "Maximum Likelihood Estimation on the Grassmannian of Lines", "domain": "science", "knowledge": "We study the positive Grassmannian through the lens of algebraic statistics. A closed formula is presented for the maximum likelihood degree of the Grassmannian of lines. We conjecture that the probability simplex contains a unique local maximum, and we present computational evidence for this.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_396ffa40b7", "k": "Bounds and Limitations on Codes Achieving List Recovery Capacity", "domain": "science", "knowledge": "In coding theory, list recoverability is a fundamental concept which robustly captures how ``spread-out'' codewords are in a code. More formally, given a code $C \\subseteq Σ^n$ and input lists $S_1, \\dots, S_n \\subseteq Σ$ of size at most $\\ell$, list recoverability requires that there are at most $L$ codewords $c \\in C$ such that $c_i \\in S_i$ for at least $(1-ρ)n$ choices of $i \\in [n]$. List recovery is an important question which has found applications in many areas, including complexity theory, property testing, compressed sensing, streaming algorithms, and cryptography. As our first main result, we establish a tight ``generalized singleton bound''. Formally, we show that for constant $\\ell, L,ρ$ and sufficiently large alphabets $Σ$, if we define $R^*=\\frac{L+1-\\ell}{L}-\\frac{L+1}{L}ρ$, it is possible for a $(ρ,\\ell,L)$ list-recoverable code to have rate $R^*-ε$ but impossible to have rate $R^*+ε$. One direction of our result already directly generalizes and improves a weaker impossibility result due to Goldberg, Shangguan, and Tamo. For our second main result, we prove that there is a fundamental shortcoming in existing methods that aim to construct explicit, optimal list-recoverable codes. Indeed, recent work has constructed explicit codes achieving list-decoding capacity (along with other related properties) using a framework introduced in the work of Alon--Edmonds--Luby (AEL). We give a meta-analysis of such constructions by presenting an ``AEL framework'' which captures all such recent constructions in the literature. Within this framework, we show that no AEL-based code can break a recently-identified list-recovery barrier for additive and linear codes.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_d7c144aac7", "k": "On the Ramsey number of a graph obtained by attaching a pendant to a path with length 1 modulo 4", "domain": "science", "knowledge": "In this paper, for odd $n$, we consider the graph $T_n$ with vertex set $\\{x_1,....,x_n,x_{n+1}\\}$ and edge set $\\{x_{1}x_{2},x_{2}x_{3},...,x_{n-1}x_{n}\\} \\cup \\{x_{\\frac{n+1}{2}}x_{n+1}\\}$ and prove that the Ramsey number $R(T_n,T_n)$ is equal to $\\frac{3n+1}{2}$ for $n \\in [1]_{4}$.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_6bf03aabff", "k": "New lower bounds for binary constant-weight codes: $A(23,6,10)\\geq 2979$ and $A(24,6,10)\\geq 4214$", "domain": "science", "knowledge": "Let $A(n,d,w)$ denote the maximum size of a binary constant-weight code of length $n$, minimum distance $d$, and weight $w$. We construct explicit codes proving $A(23,6,10)\\ge 2979$ and $A(24,6,10)\\ge 4214$. These improve the best surviving explicit codes of sizes 2969 and 4174 and surpass the corresponding 1990 bounds 2970 and 4200 of Brouwer, Shearer, Sloane and Smith, whose code listings were lost. We also obtain $A(23,6,11)\\ge 3539$ and $A(24,6,8)\\ge 1855$. All four bounds are now listed in Brouwer's online table. The constructions use a coordinate decomposition in which one half is fixed to a known code and the complementary half is selected from its full cross-compatible pool using CHILS for maximum-weight independent set. For the 2969-word $A(23,6,10)$ incumbent, exact computations with two solver families prove insertion maximality and exclude every improving exchange deleting at most three codewords. We also analyze codes invariant under prime-order permutations: several cycle types are excluded exactly, the $5+1^{18}$ type has upper bound 499, and reproducible heuristic saturation evidence is reported for the remaining types, with $13+1^{10}$ left open. Code files, an independent validator, model descriptions, and computational logs are released.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_904e42bccb", "k": "Perfect state transfer in Grover walks on normal Cayley graphs", "domain": "science", "knowledge": "A Cayley graph $\\operatorname{Cay}(Γ,S)$ over a finite group $Γ$ is said to be normal if its connection set $S$ is a union of some conjugacy classes of $Γ$. This paper investigates perfect state transfer in Grover walks on normal Cayley graphs. The Grover walk is a widely studied discrete-time quantum walk. We establish a necessary and sufficient condition for the occurrence of perfect state transfer on normal Cayley graphs. As applications, we derive explicit spectral criteria for perfect state transfer on Cayley graphs over abelian groups, dicyclic groups, and dihedral groups. These results yield several infinite families of Cayley graphs exhibiting perfect state transfer. We further obtain simple combinatorial characterizations of the existence of perfect state transfer on Cayley graphs over dihedral and dicyclic groups. Our general characterization also recovers a number of previously known results as special cases. As a further consequence, we obtain a complete characterization of perfect state transfer on unitary Cayley graphs. In particular, we prove that exactly four graphs in the class of unitary Cayley graphs exhibit perfect state transfer.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_5a9a880c05", "k": "Resolution of the ENO-TV conjecture: a parity dichotomy", "domain": "science", "knowledge": "We resolve the ENO--TV conjecture, a discrete coercivity problem in compactness theory for entropy-stable approximations of hyperbolic conservation laws. For order-$k$ essentially non-oscillatory (ENO) reconstruction from compactly supported cell averages, it asks whether the nonnegative ENO source times the $(k-1)$st power of the amplitude uniformly controls the $(k+1)$st absolute-jump moment. We prove a parity dichotomy: the estimate holds for odd $k\\ge3$ and fails for even $k\\ge4$; the known second-order case completes the classification. Localization gives a selection-independent finite-difference functional uniformly comparable to the source and reduces the conjecture to discrete interpolation. For odd orders, summation by parts reveals a hidden square; a discrete Gagliardo--Nirenberg inequality yields coercivity. For even orders, Euler-polynomial blocks from the functional's polynomial kernel yield counterexamples that persist under arbitrarily small perturbations making all affected ENO comparisons strict. We also prove two coercive estimates for every $k\\ge2$: control of jumps larger than a fixed fraction of the amplitude and of local blocks modulo sampled polynomials of degree at most $k-2$. Via the Cayley--Sylvester decomposition, we compute the dimensions of homogeneous first-cohomology spaces for the lattice shift on polynomial jump profiles. At fourth order, for a cubic flux and a globally strictly convex entropy, a total-degree-seven component of a reduced entropy-flux mismatch represents a nonzero class on profiles of degree at most two and hence has no translation-invariant finite-stencil $C^7$ local primitive at the zero constant state. Odd-order coercivity persists on globally quasi-uniform meshes, whereas for each $k\\ge2$ it fails on a fixed irregular mesh even though every interface contribution remains nonnegative. This failure is due to the mesh geometry.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_92eeb0a0c7", "k": "A Chain-Level Borsuk--Ulam Obstruction Proof of Norine's Antipodal-Coloring Conjecture", "domain": "science", "knowledge": "We prove Norine's conjecture: every red--blue edge-coloring of the \\(n\\)-dimensional hypercube \\(Q_n\\), \\(n\\geq2\\), in which antipodal edges have opposite colors contains a monochromatic path joining some vertex to its antipode. From a hypothetical counterexample we construct an antipodally equivariant, augmentation-preserving chain map from the cellular chains of the cubical boundary of a cube to subdivision-invariant polyhedral chains on a sphere of one lower dimension. A purely algebraic chain-level Borsuk--Ulam obstruction rules out this map.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_81c4f3ac7e", "k": "A complete solution to the Boots-Royle/Cao-Vince conjecture", "domain": "science", "knowledge": "Boots and Royle, and independently Cao and Vince, conjectured that the join of an edge and a path on $n-2$ vertices is the unique planar graph of maximum adjacency spectral radius for $n\\geq 9$. Tait and Tobin (JCTB, 2017) proved the conjecture for sufficiently large order. In this paper, we completely resolved the Boots-Royle/Cao-Vince conjecture.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_6957acbd58", "k": "Geometries admitting trialities for the symmetric and alternating groups", "domain": "science", "knowledge": "In incidence geometry, a triality is a symmetry cyclically exchanging triples of types of elements. Requiring geometries with trialities to satisfy standard regularity conditions makes their construction highly non trivial, and known examples are rare. In this paper, we present two infinite families of flag transitive, thin and residually connected geometries admitting trialities and no dualities with type preserving automorphism groups isomorphic to Sym(n) or Alt(n). We also develop general methods that extend to other settings. The residues of the two families are fundamentally different. In the first family, the maximal parabolics have constant size while in the other their size grows essentially linearly with the degree of the group.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_0af8241de2", "k": "Eventually nondecreasing quasi-polynomials", "domain": "science", "knowledge": "Quasi-polynomials are ubiquitous in combinatorics and algebra, as they arise in a variety of enumeration problems. Because quasi-polynomials consist of constituent polynomials, their behavior is more subtle than for a single polynomial. In particular, unlike for a polynomial, it is possible for a quasi-polynomial defined on the positive integers to have infinitely many points at which it is decreasing. In this work, we characterize quasi-polynomials of degree $d$ and period dividing $p$ that are eventually nondecreasing, i.e., that have only finitely many values at which they decrease. We then give a detailed analysis of the space of eventually nondecreasing quasi-polynomials with fixed degree $d$ and period dividing a fixed $p$ such that the $h$-vector of the quasi-polynomial is nonnegative. Using this analysis, we determine the rate of growth of the number of such quasi-polynomials as a function of the sum of the $h$-vector entries for the $0$-th constituent polynomial.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_094f9ec225", "k": "Homogeneous hypergraph regularity lemmas via $k$-strong honest definitions", "domain": "science", "knowledge": "We prove that $(k+1)$-uniform hypergraphs definable in an NIP strongly $k$-distal structure satisfy a homogeneous regularity lemma -- they can be partitioned into a bounded number of simplicial complexes, most of which are homogeneous (meaning that the restriction of the hypergraph to the simplicial complex is either complete or empty). Furthermore, the parts of the partition can be chosen uniformly definably, and the size of the partition is polynomial in the reciprocal of the error parameter. This extends the homogeneous regularity lemma proven by Chernikov and Starchenko for hypergraphs definable in a distal structure. We prove this by introducing $k$-strong honest definitions and showing that an NIP structure is strongly $k$-distal if and only if every formula $\\varphi(x_1, ..., x_k; y)$ has a $k$-strong honest definition. This extends the theory of strong honest definitions in distal structures to the higher-arity setting.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_34218815d8", "k": "On the autotopism groups and the equivalence of finite cyclic semifields", "domain": "science", "knowledge": "Special cases of finite cyclic semifields were first constructed by Hughes and Kleinfeld in 1960, and later by Sandler in 1962 and Knuth in 1965. The general construction of cyclic semifields was subsequently introduced by Petit in 1966, and later rediscovered from the perspective of irreducible semilinear transformations by Jha and Johnson in 1989. Since Sandler's foundational work in 1962, the complete determination of the autotopism groups of cyclic semifields and the full resolution of the isotopy problem for this family have remained long-standing open problems. The most significant advances in determining these autotopism groups are due to Dempwolff in 2011, who left open the case in which the field extension degree strictly divides the degree of the polynomial defining the semifield. In this paper, we provide a complete classification of cyclic semifields up to isotopy, together with the full determination of their autotopism groups, thereby closing the remaining cases left open by Dempwolff. Since cyclic semifields arise as a special instance of a broader family of maximum rank distance (MRD) codes constructed via skew polynomials, our methods also yield a complete classification of these MRD codes up to linear and semilinear equivalence over the prime field, together with an explicit description of their full automorphism groups.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_b869259ed3", "k": "On the Slice Rank of Tensors in P-Echelon Form", "domain": "science", "knowledge": "For a totally ordered finite set $A$ with the total order $Q$, a poset $P=([d],\\le_P)$, and a field $\\mathbb{F}$, a tensor $T:A^d\\longrightarrow \\mathbb{F}$ is in $P$-echelon form if $r\\le_{P}s$ in $P$ implies $a_r\\le_Q a_s$ in every tuple $(a_1,\\dots,a_d)$ in the support of $T$. We prove that if the Hasse diagram of $P$ has no isolated vertex, then tensors in $P$-echelon form with nonzero diagonal entries have full slice-rank. Our results extend and improve on recent results of Amanov and Yeliussizov.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_e132a0d928", "k": "Permutations with Verma Multiplicities $[M(p):L(q)] \\geq 2$", "domain": "science", "knowledge": "We consider permutations $q$ in the symmetric group $S_n$ whose Verma multiplicities in the principal block of $\\mathcal{O}(\\mathfrak{sl}_n)$ satisfy $[M(p):L(q)] \\geq 2$. We present a construction along with a diagrammatic visualization, showing how permutations in $S_n$ with this property generate a family of permutations in $S_{n+1}$ that share the same multiplicity property. While the method does not recover all such permutations in $S_{n+1}$, it systematically generates many new examples. In addition, we present all permutations in $S_5$ with $[M(\\mathrm{id}):L(q)] \\geq 2$ in a Bruhat diagram and describe all permutations in $S_6$ and $S_7$ with non-simple Verma multiplicities.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_d78bef3e1d", "k": "An $e$-positive classification for complete multipartite graphs", "domain": "science", "knowledge": "Shelburne and van Willigenburg (arXiv:2604.26158) characterize the Schur-positive complete multipartite graphs and leave open whether the graphs~$G=K_{(3,\\,2^β)}$ are $e$-positive. We resolve this question and, together with their classification, characterize all $e$-positive complete multipartite graphs. Our main result is an explicit, manifestly nonnegative $e$-expansion of~$X_G$ whose coefficients are expressed in terms of the restricted-injection numbers. Our main idea is to derive a marker-variable coefficient-extraction formula for the $e$-coefficients of arbitrary complete multipartite graphs from the elementary--monomial Cauchy identity. For the particular graph~$G$, this formula reduces the proof to three coefficient families, which we evaluate using Dickson polynomials and recurrences for these numbers.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_4e12bd7abf", "k": "Various Bialgebras Of Representative Functions On Free Monoids", "domain": "science", "knowledge": "Factorization and decomposition of representative functions on a free monoid X * (generated by an alphabet X ) and with values in a ring A containing Q are equivalent to factorization and decomposition of their graphs (within the A-algebra of rational noncommutative series over X ) admitting linear representations (thanks to the Kleene-Sch{ü}tzenberger theorem). To factorize and to decompose effectively these graphs, we examine various products of noncommutative series (as concatenation, shuffle and its $φ$deformations) and co-products such that, for A is a field K, their associated non graded commutative and co-noncommutative bialgebras of series are isomorphic to the Sweedler's dual of the graded noncommutative co-commutative bialgebras of polynomials having, for the concatenation, only Kleene stars of the planes as characters, or equivalently, only the planes are infinitesimal characters (thanks to a Ree's theorem like).", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_594e03b31f", "k": "Counting spanning quasi-trees of ribbon graphs: determinants and #P-completeness", "domain": "science", "knowledge": "A quasi-tree of a connected ribbon graph is a spanning ribbon subgraph with exactly one boundary component; quasi-trees play the role of spanning trees in the topological graph theory of embedded graphs. We prove that counting them is #P-complete under polynomial-time Turing reductions, already for bouquets. The proof identifies every nonempty framed chord diagram, up to natural identifications, with a 4-regular map equipped with a distinguished A-trail, in such a way that quasi-trees correspond to A-trails, whose counting is #P-complete by a theorem of Ge and Štefankovič. Through the framed Cohn-Lempel equality the count is also an interlace-polynomial evaluation - $q(H;2,1)$, the number of full-rank induced subgraphs of the looped circle graph $H$ of the diagram - placing it on the line $y=1$ left open in the complexity classification of Bläser and Hoffmann; a cloning argument then makes every fixed rational point of that line, other than the trivial $(1,1)$, #P-hard on looped circle graphs, even when a framed chord representation is supplied. On the tractable side, the same GF(2) model yields short proofs of the known determinantal cases: for orientable ribbon graphs the count is a determinant, essentially the Matrix-Quasi-tree Theorem of Merino, Moffatt and Noble, proved here via Bouchet's principal unimodularity, and for bouquets with exactly one non-orientable loop it is a sum of two orientable determinants, equivalent by a rank-one determinant identity to the determinant formula of Deng, Jin and Yan.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_97787b15fa", "k": "Cographs and Minimum Diamond-Generating Edge Sets in Boolean Lattices", "domain": "science", "knowledge": "We study a local closure operation on the cover edges of a Boolean lattice: whenever the two lower edges or the two upper edges of a square face are present, all four edges of that square are added. We prove that every set of cover edges generating the full cover graph of $\\mathcal{B}_n$ has cardinality at least $n$, and we classify all generators attaining this bound. For a graph $G$ on $[n]$, let $S_G=\\{N_G(i)\\to N_G(i)\\cup\\{i\\}:i\\in[n]\\}$. Then $S_G$ diamond-generates the full cover graph if and only if $G$ is a cograph, and every minimum-cardinality generator arises uniquely in this way. Consequently, labeled minimum diamond-generating sets of $\\mathcal{B}_n$ are in bijection with labeled cographs on $n$ vertices.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_1f68d44866", "k": "2-Connected Subgraphs of All Orders in Large Graphs with Minimum Degree at Least n/q", "domain": "science", "knowledge": "We confirm a conjecture of Liu and Ning\\cite{LiuNing}: for every fixed integer $q\\ge 3$, there exists an integer $n_0(q)$ such that every $2$-connected graph $G$ of order $n\\ge n_0(q)$ with minimum degree $δ(G)\\ge n/q$ contains a $2$-connected subgraph of order $\\ell$ for each $\\ell\\in\\{4,5,\\ldots,n\\}$. The proof has two main components. First, we construct a small $2$-connected subgraph $D$ such that every vertex outside $D$ has at least two neighbors in $D$. By successively adding the vertices outside $D$, we obtain $2$-connected subgraphs of every order from $|V(D)|$ to $n$. Second, an averaging argument on common neighborhoods yields a large complete bipartite subgraph $K_{2,t}$, which provides $2$-connected subgraphs of all the remaining orders.Finally, we propose the following conjecture: for every pair of fixed integers $r\\ge 2$ and $q\\ge 3$, every $r$-connected graph $G$ of sufficiently large order $n$ with $δ(G)\\ge n/q$ contains an $r$-connected subgraph of every order from $2r$ to $n$.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_e547d3a2c6", "k": "Stretched exponential asymptotics for bases of triangular bootstrap percolation", "domain": "science", "knowledge": "In this paper, we study a bootstrap percolation process on the finite triangular grid $\\mathfrak{T}_n$ of side length $n$. We say that a subset $η$ of points in $\\mathfrak{T}_n$ percolates if the final configuration, starting from $η$, is the whole grid $\\mathfrak{T}_n$. A basis of size $n$ is then a subset of points of $\\mathfrak{T}_n$ of minimum cardinality which percolates. In this paper, we first prove that the generating function counting bases satisfies an algebraic differential equation. Then, by analysing a modified version of this equation, we prove that the number $t_n$ of bases of size $n$ exhibits a stretched exponential asymptotic behaviour. More precisely, we show that $t_n \\sim c n!e^{\\sqrt{12n}}n^{5/12}$, for some constant $c>0$. These bases were recently shown by the second author to be in bijection with $3$-permutations avoiding the patterns $(12, 12)$ and $(231, 312)$, so this represents to our knowledge the first proven example of an asymptotic stretched exponential appearing in the study of pattern avoiding permutations.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_9399d99568", "k": "On Diophantine $m$-tuples related to primitive elements of finite fields", "domain": "science", "knowledge": "Inspired by recent works on Diophantine tuples over finite fields, in this paper we consider Diophantine tuples related to primitive elements of finite fields. Let $\\mathbb{F}_q$ be the finite field with $q$ elements and $\\mathbb{F}_q^*=\\mathbb{F}_q\\setminus\\{0\\}$ be the multiplicative cyclic group of all non-zero elements over $\\mathbb{F}_q$. An element $g\\in\\mathbb{F}_q$ is called primitive if $g$ generates the group $\\mathbb{F}_q^*$. A set $\\{x_1,x_2,\\cdots,x_m\\}\\subseteq\\mathbb{F}_q^*$ of $m$ elements is said to be a $\\mathcal{P}$-Diophantine $m$-tuple over $\\mathbb{F}_q$ if $x_ix_j+1$ is primitive for any $1\\le i\\le j\\le m$. Let $N_m$ denote the number of $\\mathcal{P}$-Diophantine tuples over $\\mathbb{F}_q$. Then we obtain the asymptotic formula $$m!\\cdot N_m=\\left(\\frac{\\varphi(q-1)}{q-1}\\right)^{m(m+1)/2}q^m+O_{m,r}\\left(q^{m-\\frac{1}{2}+r}\\right),$$ where $\\varphi(\\cdot)$ is the Euler totient function and $r\\in(0, 1/2)$ is an arbitrary real number. Moreover, we prove that there exists a $\\mathcal{P}$-Diophantine $m$-tuple over $\\mathbb{F}_q$ whenever $q\\ge \\exp(\\exp(m(m+1)))$.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_2a44f0924a", "k": "On the Computational Complexity of Local and Global Covering Numbers", "domain": "science", "knowledge": "The global and local $\\mathcal{G}$-covering number $c_{\\mathrm{g}}^{\\mathcal{G}}(H)$ and $c_{\\mathrm{l}}^{\\mathcal{G}}(H)$ encode how well the edges of a graph $H$ can be covered with graphs from a graph class $\\mathcal{G}$: in the global setting, we minimize the number of graphs from $\\mathcal{G}$ required, in the local setting how often a vertex is hit by the graphs of the cover. Within this work we consider for $\\mathcal{G}$ the graph classes $\\mathcal{B}$ of all bipartite and $\\mathcal{B}_{\\mathrm{c}}$ of all complete bipartite graphs. We give a tight lower bound on $c_{\\mathrm{l}}^{\\mathcal{B}}(H)$ in terms of the fractional chromatic number of $H$, thereby giving a local analogue of a result by Harary, Hsu and Miller. Answering a question by Fishburn and Hammer, we show that it is NP-hard to determine $c_{\\mathrm{l}}^{\\mathcal{B}_{\\mathrm{c}}}(H)$. Further, we provide a finite and monotone graph class $\\mathcal{G}$ such that $c_{\\mathrm{g}}^{\\mathcal{G}}(H)$ can be computed in constant time for every graph $H$ while determining $c_{\\mathrm{l}}^{\\mathcal{G}}(H)$ is NP-hard. This yields a natural example to a question raised by Knauer and Ueckerdt.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_25fb790e59", "k": "Dittert's conjecture in dimension 16 via a joint-deficit scaling lemma", "domain": "science", "knowledge": "Dittert's conjecture asserts that, among nonnegative $n\\times n$ matrices whose entries sum to $n$, the functional $φ(A)=\\prod_{i=1}^n r_i+\\prod_{j=1}^n c_j-\\operatorname{per}(A)$ is uniquely maximized by the uniform matrix $J_n/n$. This paper proves the conjecture for $n=16$. The key observation is that, for a near-maximizer, the deficits of the row-sum and column-sum products satisfy a single joint constraint rather than two independent bounds. Combining this joint-deficit estimate with a Pinsker-type subset-sum bound yields a sharper scalar dilation to a doubly superstochastic matrix. The Knopp-Sinkhorn boundary lower bound for permanents then excludes maximizers with a zero entry, and Hwang's positive-support theorem identifies the unique maximizer. Together with Pang's result for $n\\ge 17$ (arXiv:2606.01531), this establishes Dittert's conjecture for every $n\\ge 16$.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_711f89eeb7", "k": "A classification of locally-quasiprimitive circulant digraphs", "domain": "science", "knowledge": "Circulant digraphs are Cayley digraphs over finite cyclic groups and constitute a fundamental class of objects in algebraic graph theory. Extending the classification of locally-primitive circulant graphs \\cite{JZ-2026}, we completely determine all locally-quasiprimitive circulant digraphs. Our main theorem shows that a connected locally-quasiprimitive circulant digraph is isomorphic to one of the following: the complete graph \\(\\K_n\\), the complete bipartite graph \\(\\K_{n/2,n/2}\\), the graph \\(\\K_{n/2,n/2}-\\frac{n}{2}\\K_2\\) (with \\(n/2\\) odd), the cycle \\(\\C_n\\), the directed cycle \\(\\vec \\C_n\\), a normal circulant digraph of prime valency, the lexicographic product \\(\\vec \\C_m[\\overline{\\K_b}]\\), or the tensor product \\(\\vec \\C_m\\times \\K_b\\) with \\(\\gcd(m,b)=1\\).", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_4c7661e5b1", "k": "Polynomiality of Stretched Schubert Structure Constants and Key Coefficients", "domain": "science", "knowledge": "We prove that monomial coefficients in affine families of key and Schubert polynomials are eventually polynomial. The proof combines Demazure operators with vector partition functions and, in the Schubert case, P.~Magyar's orthodontic formula. These coefficient results extend to finite products. Using M.~Watanabe's Schubert duality, we deduce that stretched Schubert structure constants are eventually polynomial, proving a conjecture of I.~Pak and Z.~Slonim. For key polynomials, this resolves the polynomiality part of a conjecture of P.~Alexandersson and E.~Alhajjar.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_0b019b5eb2", "k": "On the possible values of the Rearrangement Number", "domain": "science", "knowledge": "The rearrangement number $\\mathfrak{rr}$ is the least cardinality of a collection of permutations of $ω$ such that every conditionally convergent real series is disrupted by some permutation in the collection. Blass, Brendle, Brian, Hamkins, Hardy, and Larson proved that $\\max\\{\\operatorname{cov}(\\mathcal N),\\mathfrak b\\}\\leq\\mathfrak{rr}\\leq\\operatorname{non}(\\mathcal M)$ and asked whether $\\mathfrak{rr}<\\mathrm{non}(\\mathcal M)$ is consistent. We prove that $\\mathfrak{rr}<\\operatorname{non}(\\mathcal M)$ is consistent with ZFC. We also prove, in a different forcing extension, that $\\max\\{\\operatorname{cov}(\\mathcal N),\\mathfrak b\\}<\\mathfrak{rr}$. We further derive consequences for the subseries number $\\mathfrak{s}_{\\mathrm{sub}}$ and the splitting number $\\mathfrak s$.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_f11d62a41e", "k": "On the toric lifting properties for simplicial $3$-spheres", "domain": "science", "knowledge": "We study the lifting problem for mod $2$ characteristic maps over simplicial $3$-spheres. Using a bad-block partition of the universal complex $X(\\mathbb{Z}_2^4)$, we prove an avoidance criterion for liftability. We show that every simplicial $3$-sphere with at most $20$ vertices has the toric lifting property. We also obtain image-size and join-type results, and prove sharpness of the image-size bound in the universal-complex sense.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_55fee7c000", "k": "A Local Classification of Four-Element Multiple Sumsets", "domain": "science", "knowledge": "For a finite set $A\\subset\\mathbb{Z}$, write $hA$ for its $h$-fold sumset, and let \\[ R(h,k)=\\{|hA|:A\\subset\\mathbb{Z},\\ |A|=k\\}. \\] We determine the part of $R(h,4)$ lying between $4h+2$ and $6h-4$: for $h=4$ the only value is $5h-1$, while for $h\\geq 5$ the only values are $5h-1$ and $5h+1$. This proves Rajagopal's conjectured gap $5h\\notin R(h,4)$ for every $h\\geq 4$. For $h\\geq 6$, it also yields the new missing interval $[5h+2,6h-4]$, which lies outside Rajagopal's general excluded set. Lev's lower bound for the successive growth of multiple sumsets reduces the problem to normalized sets of affine diameter five, of which there are only six. Reflection and four elementary exact sumset computations finish the classification.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_3bcbcc84c1", "k": "A duality of misère games and play with a pass", "domain": "science", "knowledge": "In combinatorial game theory, the choice of play convention is a fundamental aspect of the theory. The two most widely studied conventions are normal play, in which the player who makes the last move wins, and misère play, in which the player who makes the last move loses. Another well-known variant is play with a single shared pass. In such games, at most one pass may be used in total during the game: once the pass has been used, neither player may pass thereafter. Moreover, once a terminal position has been reached, passing is no longer allowed. Recall that, for misère play, one sometimes defines SG values (or Sprague-Grundy values) in the same recursive manner as in normal play but assigns the terminal position the value 1. Also, SG values can be considered for normal-play games with a pass. In this work, we propose transformations that allow both misère play and play with a pass to be treated as normal-play games. As a consequence, we show that there is a certain duality between a generalization of misère play and a generalization of play with a pass.", "authority": 4.5, "_expert": "expert_sciences"} {"id": "arxiv_061af5f7d2", "k": "Regularity and depth of binomial ideals arising from combinatorics", "domain": "science", "knowledge": "Regularity and depth of binomial ideals generated by adjacent $2$-minors together with those arising from finite lattices are studied.", "authority": 4.5, "_expert": "expert_sciences"}