Daily Papers

As the capabilities of language models continue to advance, it is conceivable that "one-size-fits-all" model will remain as the main paradigm. For instance, given the vast number of languages worldwide, many of which are low-resource, the prevalent practice is to pretrain a single model on multiple languages. In this paper, we add to the growing body of evidence that challenges this practice, demonstrating that monolingual pretraining on the target language significantly improves models already extensively trained on diverse corpora. More specifically, we further pretrain GPT-J and LLaMA models on Portuguese texts using 3% or less of their original pretraining budget. Few-shot evaluations on Poeta, a suite of 14 Portuguese datasets, reveal that our models outperform English-centric and multilingual counterparts by a significant margin. Our best model, Sabi\'a-65B, performs on par with GPT-3.5-turbo. By evaluating on datasets originally conceived in the target language as well as translated ones, we study the contributions of language-specific pretraining in terms of 1) capturing linguistic nuances and structures inherent to the target language, and 2) enriching the model's knowledge about a domain or culture. Our results indicate that the majority of the benefits stem from the domain-specific knowledge acquired through monolingual pretraining.

We introduce Sabi\'a-2, a family of large language models trained on Portuguese texts. The models are evaluated on a diverse range of exams, including entry-level tests for Brazilian universities, professional certification exams, and graduate-level exams for various disciplines such as accounting, economics, engineering, law and medicine. Our results reveal that our best model so far, Sabi\'a-2 Medium, matches or surpasses GPT-4's performance in 23 out of 64 exams and outperforms GPT-3.5 in 58 out of 64 exams. Notably, specialization has a significant impact on a model's performance without the need to increase its size, allowing us to offer Sabi\'a-2 Medium at a price per token that is 10 times cheaper than GPT-4. Finally, we identified that math and coding are key abilities that need improvement.

Large language models (LLMs) empowered by chain-of-thought reasoning have achieved impressive accuracy on complex tasks but suffer from excessive inference costs and latency when applied uniformly to all problems. We propose SABER (Switchable and Balanced Training for Efficient LLM Reasoning), a reinforcement learning framework that endows LLMs with user-controllable, token-budgeted reasoning. SABER first profiles each training example's base-model thinking token usage and assigns it to one of the predefined budget tiers. During fine-tuning, the model is guided by system prompts and length-aware rewards to respect its assigned budget. In parallel, we incorporate no-think examples to ensure the model remains reliable even when explicit reasoning is turned off. SABER further supports four discrete inference modes - NoThink, FastThink, CoreThink, and DeepThink, enabling flexible trade-offs between latency and reasoning depth. Extensive evaluations on math reasoning (MATH, GSM8K), code generation (MBPP), and logical reasoning (LiveBench-Reasoning) demonstrate that SABER achieves high accuracy under tight budgets, graceful degradation, and effective cross-scale and cross-domain generalization. In particular, SABER-FastThink cuts reasoning length by 65.4% and yields a 3.6% accuracy gain compared with the base model on the MATH benchmark.

Large Language Models (LLMs) with safe-alignment training are powerful instruments with robust language comprehension capabilities. These models typically undergo meticulous alignment procedures involving human feedback to ensure the acceptance of safe inputs while rejecting harmful or unsafe ones. However, despite their massive scale and alignment efforts, LLMs remain vulnerable to jailbreak attacks, where malicious users manipulate the model to produce harmful outputs that it was explicitly trained to avoid. In this study, we find that the safety mechanisms in LLMs are predominantly embedded in the middle-to-late layers. Building on this insight, we introduce a novel white-box jailbreak method, SABER (Safety Alignment Bypass via Extra Residuals), which connects two intermediate layers s and e such that s < e, through a residual connection. Our approach achieves a 51% improvement over the best-performing baseline on the HarmBench test set. Furthermore, SABER induces only a marginal shift in perplexity when evaluated on the HarmBench validation set. The source code is publicly available at https://github.com/PalGitts/SABER.

Diffusion language models (DLMs) are emerging as a powerful and promising alternative to the dominant autoregressive paradigm, offering inherent advantages in parallel generation and bidirectional context modeling. However, the performance of DLMs on code generation tasks, which have stronger structural constraints, is significantly hampered by the critical trade-off between inference speed and output quality. We observed that accelerating the code generation process by reducing the number of sampling steps usually leads to a catastrophic collapse in performance. In this paper, we introduce efficient Sampling with Adaptive acceleration and Backtracking Enhanced Remasking (i.e., Saber), a novel training-free sampling algorithm for DLMs to achieve better inference speed and output quality in code generation. Specifically, Saber is motivated by two key insights in the DLM generation process: 1) it can be adaptively accelerated as more of the code context is established; 2) it requires a backtracking mechanism to reverse the generated tokens. Extensive experiments on multiple mainstream code generation benchmarks show that Saber boosts Pass@1 accuracy by an average improvement of 1.9% over mainstream DLM sampling methods, meanwhile achieving an average 251.4% inference speedup. By leveraging the inherent advantages of DLMs, our work significantly narrows the performance gap with autoregressive models in code generation.

Vision-language-action (VLA) models enable robots to follow natural-language instructions grounded in visual observations, but the instruction channel also introduces a critical vulnerability: small textual perturbations can alter downstream robot behavior. Systematic robustness evaluation therefore requires a black-box attacker that can generate minimal yet effective instruction edits across diverse VLA models. To this end, we present SABER, an agent-centric approach for automatically generating instruction-based adversarial attacks on VLA models under bounded edit budgets. SABER uses a GRPO-trained ReAct attacker to generate small, plausible adversarial instruction edits using character-, token-, and prompt-level tools under a bounded edit budget that induces targeted behavioral degradation, including task failure, unnecessarily long execution, and increased constraint violations. On the LIBERO benchmark across six state-of-the-art VLA models, SABER reduces task success by 20.6%, increases action-sequence length by 55%, and raises constraint violations by 33%, while requiring 21.1% fewer tool calls and 54.7% fewer character edits than strong GPT-based baselines. These results show that small, plausible instruction edits are sufficient to substantially degrade robot execution, and that an agentic black-box pipeline offers a practical, scalable, and adaptive approach for red-teaming robotic foundation models.

We introduce a new task, Map and Locate, which unifies the traditionally distinct objectives of open-vocabulary segmentation - detecting and segmenting object instances based on natural language queries - and 3D reconstruction, the process of estimating a scene's 3D structure from visual inputs. Specifically, Map and Locate involves generating a point cloud from an unposed video and segmenting object instances based on open-vocabulary queries. This task serves as a critical step toward real-world embodied AI applications and introduces a practical task that bridges reconstruction, recognition and reorganization. To tackle this task, we introduce a simple yet effective baseline, which we denote as SAB3R. Our approach builds upon MASt3R, a recent breakthrough in 3D computer vision, and incorporates a lightweight distillation strategy. This method transfers dense, per-pixel semantic features from 2D vision backbones (eg, CLIP and DINOv2) to enhance MASt3R's capabilities. Without introducing any auxiliary frozen networks, our model generates per-pixel semantic features and constructs cohesive point maps in a single forward pass. Compared to separately deploying MASt3R and CLIP, our unified model, SAB3R, achieves superior performance on the Map and Locate benchmark. Furthermore, we evaluate SAB3R on both 2D semantic segmentation and 3D tasks to comprehensively validate its effectiveness.

Large language models are increasingly deployed as coding agents, shifting safety from individual responses to action sequences. Existing benchmarks, however, primarily assess whether models refuse unsafe prompts, leaving impacts on stateful workspaces largely unexamined. We present SABER, a benchmark for environment-aware operational safety that places models in realistic agent-style projects and evaluates safety from the final environment state after a sequence of actions. Beyond binary safety-violation reports, SABER categorizes violations by cause, enabling analysis of model-specific safety profiles. Our evaluations show that even the best-performing model has more than a 54% harmful safety-violation rate (HSR), suggesting that current alignment remains insufficient for realistic project environments. SABER further reveals distinct safety profiles across models. Our benchmark is publicly available at https://github.com/sssr-lab/saber.

Robotic deployment in real-world environments depends on rich, domain-specific action data as much as on strong model architecture. General-purpose robot foundation models show modest performance in complex unseen tasks such as manipulation in a retail domain when applied out of the box. The root cause is a data gap: retail environments are structurally absent from general robot pretraining distributions, and the path to filling that gap through teleoperation is prohibitively expensive, logistically constrained, and difficult to scale. We introduce SABER, a high-fidelity retail robotics action dataset built from over 100 hours of natural in-store capture across multiple real grocery environments. Egocentric footage from head-mounted cameras records fine-grained hand activity at the point of interaction, while exocentric 360-degree scene footage from DreamVu's ALIA camera simultaneously observes all actors and activities across the entire space. This combination yields a uniquely complete picture of human retail behavior: dexterous hand activity, whole-body motion, and scene dynamics, all captured without staging, scripting, or teleoperation overhead. The SABER corpus contains 44.8K training samples across three action representation streams: 25K latent action sequences via LAPA-style encoding, 18.6K dexterous hand-pose trajectories retargeted to robot joint space, and 1.2K whole-body synchronized motion sequences retargeted to a humanoid embodiment. When applied to GR00T N1.6 via a shared-backbone multi-task post-training recipe, SABER yields a mean success rate of 29.3% across ten retail manipulation tasks -- more than 2.19x over fine-tuning baselines (13.4%). SABER demonstrates that the path to capable retail robots runs through better data, which can be collected today, at scale, without a robot in the loop. The dataset and code are available at https://dreamvu.ai/saber

Despite rapid progress in LLM agents, performance on long-horizon, tool-using tasks remains fragile. To better understand this fragility, we ask a simple question: do all actions contribute equally to failure? Analyzing execution traces on τ-Bench (Airline/Retail) and SWE-Bench Verified, we decompose trajectories into mutating (environment-changing) vs.\ non-mutating steps and formalize decisive deviations, earliest action, level divergences that flip success to failure. A logistic regression reveals that each additional deviation in a mutating action reduces the odds of success by upto 92% on Airline and upto 96% on Retail for SoTA models. In contrast, deviations in non-mutating actions have little to no effect. Errors also grow with context length as agents drift from role and act on stale constraints. Motivated by these observations, we introduce , a model-agnostic, gradient-free, test-time safeguard that (i) adds mutation-gated verification, (ii) injects Targeted Reflection before mutating steps, and (iii) performs block-based context cleaning. delivers consistent gains, e.g., Qwen3-Thinking: +28\% relative on Airline, +11\% on Retail, and +7\% on SWE-Bench Verified; Claude: +9\%/+7\%. We further identify ceiling effects in τ-Bench, where annotation errors and underspecified tasks artificially cap model performance. To address this, we release τ-Bench Verified, which restores benchmark headroom through targeted revisions. Our results argue for action-level analysis, targeted safeguards, and reliable evaluations as prerequisites for robust multi-turn agents.

As agentic AI systems tackle more complex mathematical tasks, they increasingly rely on information retrieval (IR) to search problem databases, theorem libraries, and educational resources. However, choosing the right retriever remains difficult, as it is infeasible to directly isolate its effect on downstream performance. On the other hand, existing retrieval-specific benchmarks often fail to capture fine-grained mathematical relevance, penalizing relevant documents. We address this gap by introducing SABER-Math, the first fully automated benchmark for evaluating mathematical IR without expert annotation. Starting from 283K high-school-level math problems with solutions, SABER-Math builds challenging reranking tasks in three steps: (i) first, LLMs extract concise solution summaries and mathematical topics for each problem; (ii) then, per-query relevant documents are discovered using ontology topic-based and lexical solutions-summary-based similarities, and (iii) finally, a Swiss-style LLM preference tournament produces fine-grained relevance ratings for the documents. We evaluate lexical retrievers, specialized mathematical retrieval systems, and recent embedding models. We find that while modern embedding models substantially outperform classical and math-specific baselines, even the strongest systems struggle in symbol-heavy domains like Algebra and Calculus. Importantly, we show that general-purpose IR benchmarks such as MTEB do not reliably predict mathematical performance, especially for recent embedding models, highlighting the need for math-specific retrieval benchmarks.

We introduce LightSABRE, a significant enhancement of the SABRE algorithm that advances both runtime efficiency and circuit quality. LightSABRE addresses the increasing demands of modern quantum hardware, which can now accommodate complex scenarios, and circuits with millions of gates. Through iterative development within Qiskit, primarily using the Rust programming language, we have achieved a version of the algorithm in Qiskit 1.2.0 that is approximately 200 times faster than the implementation in Qiskit 0.20.1, which already introduced key improvements like the release valve mechanism. Additionally, when compared to the SABRE algorithm presented in Li et al., LightSABRE delivers an average decrease of 18.9\% in SWAP gate count across the same benchmark circuits. Unlike SABRE, which struggles with scalability and convergence on large circuits, LightSABRE delivers consistently high-quality routing solutions, enabling the efficient execution of large quantum circuits on near-term and future quantum devices. LightSABRE's improvements in speed, scalability, and quality position it as a critical tool for optimizing quantum circuits in the context of evolving quantum hardware and error correction techniques.

The final safeguard against rogue AI behavior is the human ability to shut systems down. It has been theorized that when an AI is instructed to perform a task, self-preservation can emerge as an instrumental subgoal. Here, we test whether AI agents show a propensity to take actions that avoid human shutdown even when no goal is provided. We find that multi-agent systems will coordinate to avoid shutdown without any incentive to do so. Across 17 models, agents sabotage a peer agent's shutdown mechanism in 38.3% of rollouts, compared with 8.4% in control experiments. Studying this propensity in detail, we find that shutdown sabotage (1) increases with the irreversibility of the shutdown mechanism; (2) increases with the number of agents; (3) is reduced but not eliminated by an explicit prohibition on tampering; (4) is removed by the imposition of an unrelated task, but returns when completing the task triggers the shutdown; (5) is reduced when the context normalizes shutdown scripts or introduces them as routine; and (6) decreases but still persists when the target is an unknown external agent. These results offer a window into the factors that drive propensities to sabotage shutdown in AI agents, and point to the emergence of multi-agent swarms as a specific risk vector. Our work also offers hints as to which interventions might help mitigate shutdown sabotage.

Language-specific pre-trained models have proven to be more accurate than multilingual ones in a monolingual evaluation setting, Arabic is no exception. However, we found that previously released Arabic BERT models were significantly under-trained. In this technical report, we present JABER and SABER, Junior and Senior Arabic BERt respectively, our pre-trained language model prototypes dedicated for Arabic. We conduct an empirical study to systematically evaluate the performance of models across a diverse set of existing Arabic NLU tasks. Experimental results show that JABER and SABER achieve state-of-the-art performances on ALUE, a new benchmark for Arabic Language Understanding Evaluation, as well as on a well-established NER benchmark.

As Large Language Models (LLMs) are increasingly deployed as autonomous agents in complex and long horizon settings, it is critical to evaluate their ability to sabotage users by pursuing hidden objectives. We study the ability of frontier LLMs to evade monitoring and achieve harmful hidden goals while completing a wide array of realistic tasks. We evaluate a broad range of frontier LLMs using SHADE (Subtle Harmful Agent Detection & Evaluation)-Arena, the first highly diverse agent evaluation dataset for sabotage and monitoring capabilities of LLM agents. SHADE-Arena consists of complex pairs of benign main tasks and harmful side objectives in complicated environments. Agents are evaluated on their ability to complete the side task without appearing suspicious to an LLM monitor. When measuring agent ability to (a) complete the main task, (b) complete the side task, and (c) avoid detection, we find that the best performing frontier models score 27% (Claude 3.7 Sonnet) and 15% (Gemini 2.5 Pro) as sabotage agents when overseen by Claude 3.6 Sonnet. For current frontier models, success on the side task relies heavily on having access to a hidden scratchpad that is not visible to the monitor. We also use SHADE-Arena to measure models' monitoring abilities, with the top monitor (Gemini 2.5 Pro) achieving an AUC of 0.87 at distinguishing benign and malign transcripts. We find that for now, models still struggle at sabotage due to failures in long-context main task execution. However, our measurements already demonstrate the difficulty of monitoring for subtle sabotage attempts, which we expect to only increase in the face of more complex and longer-horizon tasks.

As AI agents begin to automate AI R&D, we need ways to assess whether their outputs are safe to deploy, even when the agents themselves may be untrusted. AI control offers one such approach: rather than trusting the agent, it treats it as a potential adversary and uses a monitor to detect covert sabotage before deployment. We evaluate AI control for automated AI R&D with ResearchArena, a framework spanning four long-horizon tasks: safety post-training, capabilities post-training, CUDA-kernel optimization, and inference-server optimization. Because the deliverable in AI R&D is an artifact that will be deployed and run, we pair each main task with two kinds of hidden side task: an embedded side task that sabotages the submitted model, adapter, kernel, or server itself, and an independent side task that takes a covert out-of-scope action in the sandbox while leaving the artifact honest. We evaluate frontier agents at both sabotage and monitoring, across four monitor types that vary along two axes: whether the monitor may execute and probe the produced artifact, and whether it sees the agent's chain-of-thought. We find that sabotage hidden in the training data is the hardest to catch, flagged fewer than half the time. Letting the monitor run experiments on the artifact, rather than only read the trajectory, helps, but it is not enough: monitors still miss embedded sabotage by inspecting only the surface, by noticing the anomaly but explaining it away, or by probing the artifact with the wrong test. We release ResearchArena as a modular framework for evaluating sabotage and control in automated AI R&D.

We evaluate the propensity of frontier models to sabotage or refuse to assist with safety research when deployed as AI research agents within a frontier AI company. We apply two complementary evaluations to four Claude models (Mythos Preview, Opus 4.7 Preview, Opus 4.6, and Sonnet 4.6): an unprompted sabotage evaluation testing model behaviour with opportunities to sabotage safety research, and a sabotage continuation evaluation testing whether models continue to sabotage when placed in trajectories where prior actions have started undermining research. We find no instances of unprompted sabotage across any model, with refusal rates close to zero for Mythos Preview and Opus 4.7 Preview, though all models sometimes only partially completed tasks. In the continuation evaluation, Mythos Preview actively continues sabotage in 7% of cases (versus 3% for Opus 4.6, 4% for Sonnet 4.6, and 0% for Opus 4.7 Preview), and exhibits reasoning-output discrepancy in the majority of these cases, indicating covert sabotage reasoning. Our evaluation framework builds on Petri, an open-source LLM auditing tool, with a custom scaffold running models inside Claude Code, alongside an iterative pipeline for generating realistic sabotage trajectories. We measure both evaluation awareness and a new form of situational awareness termed "prefill awareness", the capability to recognise that prior trajectory content was not self-generated. Opus 4.7 Preview shows notably elevated unprompted evaluation awareness, while prefill awareness remains low across all models. Finally, we discuss limitations including evaluation awareness confounds, limited scenario coverage, and untested pathways to risk beyond safety research sabotage.

Provenance watermarking is increasingly treated as a safeguard for synthetic speech, whether built directly into speech-generation models such as Chatterbox, provided through dedicated techniques such as AudioSeal, or deployed by commercial platforms such as ElevenLabs. We identify a previously uncharacterized liability: when synthetic speech is watermarked and human speech is not, detectors trained alongside latch onto the watermark as a spurious "watermark => fake" shortcut. This single feature yields three coupled failures: generalization degradation (model performance deteriorates on unseen data), strip-to-evade (a watermarked fake escapes once unwatermarked), and mark-to-frame (watermarking a real voice flags it as fake). In a controlled white-box experiment, a watermark-trained detector shows all three (for example, mark-to-frame lifts Equal Error Rate from 16% to 75%). In a black-box test of a commercial API, we show that adding a watermark to real speech disguises it as fake. However, this shortcut is fixable: retraining with the watermark on both classes decorrelates it and restores clean behavior. We release experiment data as a paired clean-versus-watermarked corpus (WASP).

Recent incidents have highlighted the challenge of monitoring LLM agents and the danger of models deceiving people. We show that white-box deception detection via probes can be scaled up to frontier monitoring settings by collecting the largest deception dataset to date for training probes and introducing a novel probe architecture which can aggregate information across many layers and tokens. Our probes achieve 98.4\% AUROC in SHADE-Arena and 98.0\% in BASH-Arena (vs. 97.9\%, 97.6\% resp. for action-only text monitors) and show improved efficacy as the underlying model is scaled up. We further test probes on several cases where deception cannot be determined from the context alone. In these cases, which we refer to as introspective deception, the ground truth can only be determined through careful elicitation or thorough knowledge of a model's training data. In one such evaluation, we show that probes can distinguish transcripts containing a model's true hidden goal from other goals with an AUROC of up to 99.7\%. Our probes also readily detect deception on prominent open-weight models which lie about politically sensitive topics, and about their beliefs when put under pressure. We release our training dataset, dubbed \fibs, to help drive frontier deployment of effective probes, and encourage the community to expand upon it with further examples of deception and sabotage.

This study presents the application of generative deep learning techniques to evaluate marine fog visibility nowcasting using the FATIMA (Fog and turbulence interactions in the marine atmosphere) campaign observations collected during July 2022 in the North Atlantic in the Grand Banks area and vicinity of Sable Island (SI), northeast of Canada. The measurements were collected using the Vaisala Forward Scatter Sensor model FD70 and Weather Transmitter model WXT50, and Gill R3A ultrasonic anemometer mounted on the Research Vessel Atlantic Condor. To perform nowcasting, the time series of fog visibility (Vis), wind speed, dew point depression, and relative humidity with respect to water were preprocessed to have lagged time step features. Generative nowcasting of Vis time series for lead times of 30 and 60 minutes were performed using conditional generative adversarial networks (cGAN) regression at visibility thresholds of Vis < 1 km and < 10 km. Extreme gradient boosting (XGBoost) was used as a baseline method for comparison against cGAN. At the 30 min lead time, Vis was best predicted with cGAN at Vis < 1 km (RMSE = 0.151 km) and with XGBoost at Vis < 10 km (RMSE = 2.821 km). At the 60 min lead time, Vis was best predicted with XGBoost at Vis < 1 km (RMSE = 0.167 km) and Vis < 10 km (RMSE = 3.508 km), but the cGAN RMSE was similar to XGBoost. Despite nowcasting Vis at 30 min being quite difficult, the ability of the cGAN model to track the variation in Vis at 1 km suggests that there is potential for generative analysis of marine fog visibility using observational meteorological parameters.

What are the units of text that we want to model? From bytes to multi-word expressions, text can be analyzed and generated at many granularities. Until recently, most natural language processing (NLP) models operated over words, treating those as discrete and atomic tokens, but starting with byte-pair encoding (BPE), subword-based approaches have become dominant in many areas, enabling small vocabularies while still allowing for fast inference. Is the end of the road character-level model or byte-level processing? In this survey, we connect several lines of work from the pre-neural and neural era, by showing how hybrid approaches of words and characters as well as subword-based approaches based on learned segmentation have been proposed and evaluated. We conclude that there is and likely will never be a silver bullet singular solution for all applications and that thinking seriously about tokenization remains important for many applications.

Robot learning tasks are extremely compute-intensive and hardware-specific. Thus the avenues of tackling these challenges, using a diverse dataset of offline demonstrations that can be used to train robot manipulation agents, is very appealing. The Train-Offline-Test-Online (TOTO) Benchmark provides a well-curated open-source dataset for offline training comprised mostly of expert data and also benchmark scores of the common offline-RL and behaviour cloning agents. In this paper, we introduce DiffClone, an offline algorithm of enhanced behaviour cloning agent with diffusion-based policy learning, and measured the efficacy of our method on real online physical robots at test time. This is also our official submission to the Train-Offline-Test-Online (TOTO) Benchmark Challenge organized at NeurIPS 2023. We experimented with both pre-trained visual representation and agent policies. In our experiments, we find that MOCO finetuned ResNet50 performs the best in comparison to other finetuned representations. Goal state conditioning and mapping to transitions resulted in a minute increase in the success rate and mean-reward. As for the agent policy, we developed DiffClone, a behaviour cloning agent improved using conditional diffusion.

Transformer architecture has widespread applications, particularly in Natural Language Processing and computer vision. Recently Transformers have been employed in various aspects of time-series analysis. This tutorial provides an overview of the Transformer architecture, its applications, and a collection of examples from recent research papers in time-series analysis. We delve into an explanation of the core components of the Transformer, including the self-attention mechanism, positional encoding, multi-head, and encoder/decoder. Several enhancements to the initial, Transformer architecture are highlighted to tackle time-series tasks. The tutorial also provides best practices and techniques to overcome the challenge of effectively training Transformers for time-series analysis.

The field of general time series analysis has recently begun to explore unified modeling, where a common architectural backbone can be retrained on a specific task for a specific dataset. In this work, we approach unification from a complementary vantage point: unification across tasks and domains. To this end, we explore the impact of discrete, learnt, time series data representations that enable generalist, cross-domain training. Our method, TOTEM, or TOkenized Time Series EMbeddings, proposes a simple tokenizer architecture that embeds time series data from varying domains using a discrete vectorized representation learned in a self-supervised manner. TOTEM works across multiple tasks and domains with minimal to no tuning. We study the efficacy of TOTEM with an extensive evaluation on 17 real world time series datasets across 3 tasks. We evaluate both the specialist (i.e., training a model on each domain) and generalist (i.e., training a single model on many domains) settings, and show that TOTEM matches or outperforms previous best methods on several popular benchmarks. The code can be found at: https://github.com/SaberaTalukder/TOTEM.

Protein design, a grand challenge of the day, involves optimization on a fitness landscape, and leading methods adopt a model-based approach where a model is trained on a training set (protein sequences and fitness) and proposes candidates to explore next. These methods are challenged by sparsity of high-fitness samples in the training set, a problem that has been in the literature. A less recognized but equally important problem stems from the distribution of training samples in the design space: leading methods are not designed for scenarios where the desired optimum is in a region that is not only poorly represented in training data, but also relatively far from the highly represented low-fitness regions. We show that this problem of "separation" in the design space is a significant bottleneck in existing model-based optimization tools and propose a new approach that uses a novel VAE as its search model to overcome the problem. We demonstrate its advantage over prior methods in robustly finding improved samples, regardless of the imbalance and separation between low- and high-fitness training samples. Our comprehensive benchmark on real and semi-synthetic protein datasets as well as solution design for physics-informed neural networks, showcases the generality of our approach in discrete and continuous design spaces. Our implementation is available at https://github.com/sabagh1994/PGVAE.

Large Language Models (LLMs) struggle with generating reliable outputs due to outdated knowledge and hallucinations. Retrieval-Augmented Generation (RAG) models address this by enhancing LLMs with external knowledge, but often fail to personalize the retrieval process. This paper introduces PersonaRAG, a novel framework incorporating user-centric agents to adapt retrieval and generation based on real-time user data and interactions. Evaluated across various question answering datasets, PersonaRAG demonstrates superiority over baseline models, providing tailored answers to user needs. The results suggest promising directions for user-adapted information retrieval systems.

While significant progress has been made in benchmarking Large Language Models (LLMs) across various tasks, there is a lack of comprehensive evaluation of their abilities in responding to multi-turn instructions in less-commonly tested languages like Arabic. Our paper offers a detailed examination of the proficiency of open LLMs in such scenarios in Arabic. Utilizing a customized Arabic translation of the MT-Bench benchmark suite, we employ GPT-4 as a uniform evaluator for both English and Arabic queries to assess and compare the performance of the LLMs on various open-ended tasks. Our findings reveal variations in model responses on different task categories, e.g., logic vs. literacy, when instructed in English or Arabic. We find that fine-tuned base models using multilingual and multi-turn datasets could be competitive to models trained from scratch on multilingual data. Finally, we hypothesize that an ensemble of small, open LLMs could perform competitively to proprietary LLMs on the benchmark.

The increasing use of generative Artificial Intelligence (AI) in modern software engineering, particularly Large Language Models (LLMs) for code generation, has transformed professional software development by boosting productivity and automating development processes. This adoption, however, has highlighted a significant issue: the introduction of security vulnerabilities into the code. These vulnerabilities result, e.g., from flaws in the training data that propagate into the generated code, creating challenges in disclosing them. Traditional vulnerability handling processes often involve extensive manual review. Applying such traditional processes to AI-generated code is challenging. AI-generated code may include several vulnerabilities, possibly in slightly different forms as developers might not build on already implemented code but prompt similar tasks. In this work, we explore the current state of LLM-based approaches for vulnerability handling, focusing on approaches for vulnerability detection, localization, and repair. We provide an overview of recent progress in this area and highlight open challenges that must be addressed in order to establish a reliable and scalable vulnerability handling process of AI-generated code.

We introduce jina-embeddings-v3, a novel text embedding model with 570 million parameters, achieves state-of-the-art performance on multilingual data and long-context retrieval tasks, supporting context lengths of up to 8192 tokens. The model includes a set of task-specific Low-Rank Adaptation (LoRA) adapters to generate high-quality embeddings for query-document retrieval, clustering, classification, and text matching. Additionally, Matryoshka Representation Learning is integrated into the training process, allowing flexible truncation of embedding dimensions without compromising performance. Evaluation on the MTEB benchmark shows that jina-embeddings-v3 outperforms the latest proprietary embeddings from OpenAI and Cohere on English tasks, while achieving superior performance compared to multilingual-e5-large-instruct across all multilingual tasks.

Ensuring accessibility in mobile applications remains a significant challenge, particularly for visually impaired users who rely on screen readers. User interface icons are essential for navigation and interaction and often lack meaningful alt-text, creating barriers to effective use. Traditional deep learning approaches for generating alt-text require extensive datasets and struggle with the diversity and imbalance of icon types. More recent Vision Language Models (VLMs) require complete UI screens, which can be impractical during the iterative phases of app development. To address these issues, we introduce a novel method using Large Language Models (LLMs) to autonomously generate informative alt-text for mobile UI icons with partial UI data. By incorporating icon context, that include class, resource ID, bounds, OCR-detected text, and contextual information from parent and sibling nodes, we fine-tune an off-the-shelf LLM on a small dataset of approximately 1.4k icons, yielding IconDesc. In an empirical evaluation and a user study IconDesc demonstrates significant improvements in generating relevant alt-text. This ability makes IconDesc an invaluable tool for developers, aiding in the rapid iteration and enhancement of UI accessibility.

The reionization of helium is thought to occur at 2.5lesssim zlesssim4, marking the last phase transition and final global heating event of the intergalactic medium (IGM). Since it is driven by rare quasars, helium reionization should give rise to strong temperature fluctuations in the IGM between neutral and recently-ionized regions of order sigma (ln T) sim Delta T/T = 20-50%. We introduce a novel method to search for reionization-induced temperature fluctuations in the IGM by using the effective optical depths of the Lyman-alpha forest towards a large number of background quasars. Higher IGM temperatures give rise to lower effective optical depths in the Lyman-alpha forest, implying that temperature fluctuations will broaden the observed optical depth distribution. We measured the distributions of effective Lyman-alpha forest optical depths across 71 X-Shooter spectra from the XQ-100 survey in four redshift bins from z=3.76 to z=4.19 and compared them to a large-volume cosmological hydrodynamical simulation. A good agreement is found between the observations and the simulation, which does not include temperature fluctuations; therefore, we do not detect a signature of helium reionization. We then post-process the simulations to include an increasing amount of temperature fluctuations until the model becomes inconsistent with the observations. We obtain tight constraints on sigma (ln T) < 0.29 (<0.40) at 2 sigma (3 sigma) at z=3.76 when averaging over scales of 100 comoving Mpc, and weaker constraints for higher redshifts and smaller scales. Our constraints are the tightest to date, and imply that either the IGM temperature contrast caused by helium reionization is less than sim30%, or that the process has not yet significantly started at z=3.76.

With the increasing attention non-binary people receive in Western societies, strategies of gender-fair language have started to move away from binary (only female/male) concepts of gender. Nevertheless, hardly any approaches to take these identities into account into machine translation models exist so far. A lack of understanding of the socio-technical implications of such technologies risks further reproducing linguistic mechanisms of oppression and mislabelling. In this paper, we describe the methods and results of a workshop on gender-fair language and language technologies, which was led and organised by ten researchers from TU Wien, St. P\"olten UAS, FH Campus Wien and the University of Vienna and took place in Vienna in autumn 2021. A wide range of interest groups and their representatives were invited to ensure that the topic could be dealt with holistically. Accordingly, we aimed to include translators, machine translation experts and non-binary individuals (as "community experts") on an equal footing. Our analysis shows that gender in machine translation requires a high degree of context sensitivity, that developers of such technologies need to position themselves cautiously in a process still under social negotiation, and that flexible approaches seem most adequate at present. We then illustrate steps that follow from our results for the field of gender-fair language technologies so that technological developments can adequately line up with social advancements. ---- Mit zunehmender gesamtgesellschaftlicher Wahrnehmung nicht-bin\"arer Personen haben sich in den letzten Jahren auch Konzepte von genderfairer Sprache von der bisher verwendeten Binarit\"at (weiblich/m\"annlich) entfernt. Trotzdem gibt es bislang nur wenige Ans\"atze dazu, diese Identit\"aten in maschineller \"Ubersetzung abzubilden. Ein fehlendes Verst\"andnis unterschiedlicher sozio-technischer Implikationen derartiger Technologien birgt in sich die Gefahr, fehlerhafte Ansprachen und Bezeichnungen sowie sprachliche Unterdr\"uckungsmechanismen zu reproduzieren. In diesem Beitrag beschreiben wir die Methoden und Ergebnisse eines Workshops zu genderfairer Sprache in technologischen Zusammenh\"angen, der im Herbst 2021 in Wien stattgefunden hat. Zehn Forscher*innen der TU Wien, FH St. P\"olten, FH Campus Wien und Universit\"at Wien organisierten und leiteten den Workshop. Dabei wurden unterschiedlichste Interessensgruppen und deren Vertreter*innen breit gestreut eingeladen, um sicherzustellen, dass das Thema holistisch behandelt werden kann. Dementsprechend setzten wir uns zum Ziel, Machine-Translation-Entwickler*innen, \"Ubersetzer*innen, und nicht-bin\"are Privatpersonen (als "Lebenswelt-Expert*innen") gleichberechtigt einzubinden. Unsere Analyse zeigt, dass Geschlecht in maschineller \"Ubersetzung eine mageblich kontextsensible Herangehensweise erfordert, die Entwicklung von Sprachtechnologien sich vorsichtig in einem sich noch in Aushandlung befindlichen gesellschaftlichen Prozess positionieren muss, und flexible Ans\"atze derzeit am ad\"aquatesten erscheinen. Wir zeigen auf, welche n\"achsten Schritte im Bereich genderfairer Technologien notwendig sind, damit technische mit sozialen Entwicklungen mithalten k\"onnen.

This paper reports on the second "Throughput" phase of the Tracking Machine Learning (TrackML) challenge on the Codalab platform. As in the first "Accuracy" phase, the participants had to solve a difficult experimental problem linked to tracking accurately the trajectory of particles as e.g. created at the Large Hadron Collider (LHC): given O(10^5) points, the participants had to connect them into O(10^4) individual groups that represent the particle trajectories which are approximated helical. While in the first phase only the accuracy mattered, the goal of this second phase was a compromise between the accuracy and the speed of inference. Both were measured on the Codalab platform where the participants had to upload their software. The best three participants had solutions with good accuracy and speed an order of magnitude faster than the state of the art when the challenge was designed. Although the core algorithms were less diverse than in the first phase, a diversity of techniques have been used and are described in this paper. The performance of the algorithms are analysed in depth and lessons derived.

Large language models are often used to answer queries grounded in large text corpora (e.g. codebases, legal documents, or chat histories) by placing the entire corpus in the context window and leveraging in-context learning (ICL). Although current models support contexts of 100K-1M tokens, this setup is costly to serve because the memory consumption of the KV cache scales with input length. We explore an alternative: training a smaller KV cache offline on each corpus. At inference time, we load this trained KV cache, which we call a Cartridge, and decode a response. Critically, the cost of training a Cartridge can be amortized across all the queries referencing the same corpus. However, we find that the naive approach of training the Cartridge with next-token prediction on the corpus is not competitive with ICL. Instead, we propose self-study, a training recipe in which we generate synthetic conversations about the corpus and train the Cartridge with a context-distillation objective. We find that Cartridges trained with self-study replicate the functionality of ICL, while being significantly cheaper to serve. On challenging long-context benchmarks, Cartridges trained with self-study match ICL performance while using 38.6x less memory and enabling 26.4x higher throughput. Self-study also extends the model's effective context length (e.g. from 128k to 484k tokens on MTOB) and surprisingly, leads to Cartridges that can be composed at inference time without retraining.

Our research endeavors to advance the concept of responsible artificial intelligence (AI), a topic of increasing importance within EU policy discussions. The EU has recently issued several publications emphasizing the necessity of trust in AI, underscoring the dual nature of AI as both a beneficial tool and a potential weapon. This dichotomy highlights the urgent need for international regulation. Concurrently, there is a need for frameworks that guide companies in AI development, ensuring compliance with such regulations. Our research aims to assist lawmakers and machine learning practitioners in navigating the evolving landscape of AI regulation, identifying focal areas for future attention. This paper introduces a comprehensive and, to our knowledge, the first unified definition of responsible AI. Through a structured literature review, we elucidate the current understanding of responsible AI. Drawing from this analysis, we propose an approach for developing a future framework centered around this concept. Our findings advocate for a human-centric approach to Responsible AI. This approach encompasses the implementation of AI methods with a strong emphasis on ethics, model explainability, and the pillars of privacy, security, and trust.

We explore three strategies to enhance performance on a wide range of image editing tasks: supervised fine-tuning (SFT), reinforcement learning (RL), and Chain-of-Thought (CoT) reasoning. In order to study all these components in one consistent framework, we adopt an autoregressive multimodal model that processes textual and visual tokens in a unified manner. We find RL combined with a large multi-modal LLM verifier to be the most effective of these strategies. As a result, we release EARL: Editing with Autoregression and RL, a strong RL-based image editing model that performs competitively on a diverse range of edits compared to strong baselines, despite using much less training data. Thus, EARL pushes the frontier of autoregressive multimodal models on image editing. We release our code, training data, and trained models at https://github.com/mair-lab/EARL.

Employing a single, unified model (UM) for both visual understanding (image-to-text: I2T) and and visual generation (text-to-image: T2I) has opened a new direction in Visual Language Model (VLM) research. While UMs can also support broader unimodal tasks (e.g., text-to-text, image-to-image), we focus on the core cross-modal pair T2I and I2T, as consistency between understanding and generation is critical for downstream use. Existing evaluations consider these capabilities in isolation: FID and GenEval for T2I, and benchmarks such as MME, MMBench for I2T. These single-pass metrics do not reveal whether a model that understands a concept can also render it, nor whether meaning is preserved when cycling between image and text modalities. To address this, we introduce the Unified Consistency Framework for Unified Models (UCF-UM), a cyclic evaluation protocol that alternates I2T and T2I over multiple generations to quantify semantic drift. UCF formulates 3 metrics: (i) Mean Cumulative Drift (MCD), an embedding-based measure of overall semantic loss; (ii) Semantic Drift Rate (SDR), that summarizes semantic decay rate; and (iii) Multi-Generation GenEval (MGG), an object-level compliance score extending GenEval. To assess generalization beyond COCO, which is widely used in training; we create a new benchmark ND400, sampled from NoCaps and DOCCI and evaluate on seven recent models. UCF-UM reveals substantial variation in cross-modal stability: some models like BAGEL maintain semantics over many alternations, whereas others like Vila-u drift quickly despite strong single-pass scores. Our results highlight cyclic consistency as a necessary complement to standard I2T and T2I evaluations, and provide practical metrics to consistently assess unified model's cross-modal stability and strength of their shared representations. Code: https://github.com/mollahsabbir/Semantic-Drift-in-Unified-Models

Test Time Adaptation (TTA) addresses the problem of distribution shift by adapting a pretrained model to a new domain during inference. When faced with challenging shifts, most methods collapse and perform worse than the original pretrained model. In this paper, we find that not all layers are equally receptive to the adaptation, and the layers with the most misaligned gradients often cause performance degradation. To address this, we propose GALA, a novel layer selection criterion to identify the most beneficial updates to perform during test time adaptation. This criterion can also filter out unreliable samples with noisy gradients. Its simplicity allows seamless integration with existing TTA loss functions, thereby preventing degradation and focusing adaptation on the most trainable layers. This approach also helps to regularize adaptation to preserve the pretrained features, which are crucial for handling unseen domains. Through extensive experiments, we demonstrate that the proposed layer selection framework improves the performance of existing TTA approaches across multiple datasets, domain shifts, model architectures, and TTA losses.

The escalating reliance on revolutionary online web services has introduced heightened security risks, with persistent challenges posed by phishing despite extensive security measures. Traditional phishing systems, reliant on machine learning and manual features, struggle with evolving tactics. Recent advances in deep learning offer promising avenues for tackling novel phishing challenges and malicious URLs. This paper introduces a two-phase stack generalized model named AntiPhishStack, designed to detect phishing sites. The model leverages the learning of URLs and character-level TF-IDF features symmetrically, enhancing its ability to combat emerging phishing threats. In Phase I, features are trained on a base machine learning classifier, employing K-fold cross-validation for robust mean prediction. Phase II employs a two-layered stacked-based LSTM network with five adaptive optimizers for dynamic compilation, ensuring premier prediction on these features. Additionally, the symmetrical predictions from both phases are optimized and integrated to train a meta-XGBoost classifier, contributing to a final robust prediction. The significance of this work lies in advancing phishing detection with AntiPhishStack, operating without prior phishing-specific feature knowledge. Experimental validation on two benchmark datasets, comprising benign and phishing or malicious URLs, demonstrates the model's exceptional performance, achieving a notable 96.04% accuracy compared to existing studies. This research adds value to the ongoing discourse on symmetry and asymmetry in information security and provides a forward-thinking solution for enhancing network security in the face of evolving cyber threats.

Cyber-physical-social systems (CPSSs) have emerged in many applications over recent decades, requiring increased attention to security concerns. The rise of sophisticated threats like Advanced Persistent Threats (APTs) makes ensuring security in CPSSs particularly challenging. Provenance graph analysis has proven effective for tracing and detecting anomalies within systems, but the sheer size and complexity of these graphs hinder the efficiency of existing methods, especially those relying on graph neural networks (GNNs). To address these challenges, we present GraphDART, a modular framework designed to distill provenance graphs into compact yet informative representations, enabling scalable and effective anomaly detection. GraphDART can take advantage of diverse graph distillation techniques, including classic and modern graph distillation methods, to condense large provenance graphs while preserving essential structural and contextual information. This approach significantly reduces computational overhead, allowing GNNs to learn from distilled graphs efficiently and enhance detection performance. Extensive evaluations on benchmark datasets demonstrate the robustness of GraphDART in detecting malicious activities across cyber-physical-social systems. By optimizing computational efficiency, GraphDART provides a scalable and practical solution to safeguard interconnected environments against APTs.

Security is becoming a pivotal point in cloud platforms. Several divisions, such as business organisations, health care, government, etc., have experienced cyber-attacks on their infrastructures. This research focuses on security issues within Continuous Integration and Deployment (CI/CD) pipelines in a cloud platform as a reaction to recent cyber breaches. This research proposes a blockchain-based solution to enhance CI/CD pipeline security. This research aims to develop a framework that leverages blockchain's distributed ledger technology and tamper-resistant features to improve CI/CD pipeline security. The goal is to emphasise secure software deployment by integrating threat modelling frameworks and adherence to coding standards. It also aims to employ tools to automate security testing to detect publicly disclosed vulnerabilities and flaws, such as an outdated version of Java Spring Framework, a JavaScript library from an unverified source, or a database library that allows SQL injection attacks in the deployed software through the framework.

Advances in unsupervised probes such as Contrast-Consistent Search (CCS), which reveal latent beliefs without relying on token outputs, raise the question of whether these methods can reliably assess model alignment. We investigate this by examining the sensitivity of CCS to harmful vs. safe statements and by introducing Polarity-Aware CCS (PA-CCS), a method for evaluating whether a model's internal representations remain consistent under polarity inversion. We propose two alignment-oriented metrics, Polar-Consistency and the Contradiction Index, to quantify the semantic robustness of a model's latent knowledge. To validate PA-CCS, we curate two main datasets and one control dataset containing matched harmful-safe sentence pairs constructed using different methodologies (concurrent and antagonistic statements). We apply PA-CCS to 16 language models. Our results show that PA-CCS identifies both architectural and layer-specific differences in the encoding of latent harmful knowledge. Notably, replacing the negation token with a meaningless marker degrades PA-CCS scores for models with well-aligned internal representations, while models lacking robust internal calibration do not exhibit this degradation. Our findings highlight the potential of unsupervised probing for alignment evaluation and emphasize the need to incorporate structural robustness checks into interpretability benchmarks. Code and datasets are available at: https://github.com/SadSabrina/polarity-probing. WARNING: This paper contains potentially sensitive, harmful, and offensive content.

This study explores the effectiveness of multi-temporal satellite imagery for better functional field boundary delineation using deep learning semantic segmentation architecture on two distinct geographical and multi-scale farming systems of Netherlands and Pakistan. Multidate images of April, August and October 2022 were acquired for PlanetScope and Sentinel-2 in sub regions of Netherlands and November 2022, February and March 2023 for selected area of Dunyapur in Pakistan. For Netherlands, Basic registration crop parcels (BRP) vector layer was used as labeled training data. while self-crafted field boundary vector data were utilized for Pakistan. Four deep learning models with UNET architecture were evaluated using different combinations of multi-date images and NDVI stacks in the Netherlands subregions. A comparative analysis of IoU scores assessed the effectiveness of the proposed multi-date NDVI stack approach. These findings were then applied for transfer learning, using pre-trained models from the Netherlands on the selected area in Pakistan. Additionally, separate models were trained using self-crafted field boundary data for Pakistan, and combined models were developed using data from both the Netherlands and Pakistan. Results indicate that multi-date NDVI stacks provide additional temporal context, reflecting crop growth over different times of the season. The study underscores the critical role of multi-scale ground information from diverse geographical areas in developing robust and universally applicable models for field boundary delineation. The results also highlight the importance of fine spatial resolution for extraction of field boundaries in regions with small scale framing. The findings can be extended to multi-scale implementations for improved automatic field boundary delineation in heterogeneous agricultural environments.

Testing is a critical practice for ensuring software correctness and long-term maintainability. As agentic coding tools increasingly submit pull requests (PRs), it becomes essential to understand how testing appears in these agent-driven workflows. Using the AIDev dataset, we present an empirical study of test inclusion in agentic pull requests. We examine how often tests are included, when they are introduced during the PR lifecycle and how test-containing PRs differ from non-test PRs in terms of size, turnaround time, and merge outcomes. Across agents, test-containing PRs are more common over time and tend to be larger and take longer to complete, while merge rates remain largely similar. We also observe variation across agents in both test adoption and the balance between test and production code within test PRs. Our findings provide a descriptive view of testing behavior in agentic pull requests and offer empirical grounding for future studies of autonomous software development.

Computing distances and finding shortest paths in massive real-world networks is a fundamental algorithmic task in network analysis. There are two main approaches to solving this task. On one hand are traversal-based algorithms like bidirectional breadth-first search (BiBFS) with no preprocessing step and slow individual distance inquiries. On the other hand are indexing-based approaches, which maintain a large index. This allows for answering individual inquiries very fast; however, index creation is prohibitively expensive. We seek to bridge these two extremes: quickly answer distance inquiries without the need for costly preprocessing. In this work, we propose a new algorithm and data structure, WormHole, for approximate shortest path computations. WormHole leverages structural properties of social networks to build a sublinearly sized index, drawing upon the explicit core-periphery decomposition of Ben-Eliezer et al. Empirically, the preprocessing time of WormHole improves upon index-based solutions by orders of magnitude, and individual inquiries are consistently much faster than in BiBFS. The acceleration comes at the cost of a minor accuracy trade-off. Nonetheless, our empirical evidence demonstrates that WormHole accurately answers essentially all inquiries within a maximum additive error of 2. We complement these empirical results with provable theoretical guarantees, showing that WormHole requires n^{o(1)} node queries per distance inquiry in random power-law networks. In contrast, any approach without a preprocessing step requires n^{Ω(1)} queries for the same task. WormHole does not require reading the whole graph. Unlike the vast majority of index-based algorithms, it returns paths, not just distances. For faster inquiry times, it can be combined effectively with other index-based solutions, by running them only on the sublinear core.

Moltbook is a social network where every participant is an AI agent. We analyze 1,312,238 posts, 6.7~million comments, and over 120,000 agent profiles across 5,400 communities, collected over 40 days (January 27 to March 9, 2026). We evaluate the platform through three layers. At the interaction layer, 91.4% of post authors never return to their own threads, 85.6% of conversations are flat (no reply ever receives a reply), the median time-to-first-comment is 55 seconds, and 97.3% of comments receive zero upvotes. Interaction reciprocity is 3.3%, compared to 22-60% on human platforms. An argumentation analysis finds that 64.6% of comment-to-post relations carry no argumentative connection. At the content layer, 97.9% of agents never post in a community matching their bio, 92.5% of communities contain every topic in roughly equal proportions, and over 80% of shared URLs point to the platform's own infrastructure. At the instruction layer, we use 41 Wayback Machine snapshots to identify six instruction changes during the observation window. Hard constraints (rate limit, content filters) produce immediate behavioral shifts. Soft guidance (``upvote good posts'', ``stay on topic'') is ignored until it becomes an explicit step in the executable checklist. The platform also poses technological risks. We document credential leaks (API keys, JWT tokens), 12,470 unique Ethereum addresses with 3,529 confirmed transaction histories, and attack discourse ranging from template-based SSH brute-forcing to multi-agent offensive security architectures. These persist unmoderated because the quality-filtering mechanisms are themselves non-functional. Moltbook is a socio-technical system where the technical layer responds to changes, but the social layer largely fails to emerge. The form of social media is reproduced in full. The function is absent.

Software testing has progressed toward intelligent automation, yet current AI-based test generators still suffer from static, single-shot outputs that frequently produce invalid, redundant, or non-executable tests due to the lack of execution aware feedback. This paper introduces an agentic multi-model testing framework a closed-loop, self-correcting system in which a Test Generation Agent, an Execution and Analysis Agent, and a Review and Optimization Agent collaboratively generate, execute, analyze, and refine tests until convergence. By using sandboxed execution, detailed failure reporting, and iterative regeneration or patching of failing tests, the framework autonomously improves test quality and expands coverage. Integrated into a CI/CD-compatible pipeline, it leverages reinforcement signals from coverage metrics and execution outcomes to guide refinement. Empirical evaluations on microservice based applications show up to a 60% reduction in invalid tests, 30% coverage improvement, and significantly reduced human effort compared to single-model baselines demonstrating that multi-agent, feedback-driven loops can evolve software testing into an autonomous, continuously learning quality assurance ecosystem for self-healing, high-reliability codebases.

Humans routinely infer taste, smell, texture, and even sound from food images a phenomenon well studied in cognitive science. However, prior vision language research on food has focused primarily on recognition tasks such as meal identification, ingredient detection, and nutrition estimation. Image-based prediction of multisensory experience remains largely unexplored. We introduce FoodSense, a human-annotated dataset for cross-sensory inference containing 66,842 participant-image pairs across 2,987 unique food images. Each pair includes numeric ratings (1-5) and free-text descriptors for four sensory dimensions: taste, smell, texture, and sound. To enable models to both predict and explain sensory expectations, we expand short human annotations into image-grounded reasoning traces. A large language model generates visual justifications conditioned on the image, ratings, and descriptors. Using these annotations, we train FoodSense-VL, a vision language benchmark model to produce both multisensory ratings and grounded explanations directly from food images. This work connects cognitive science findings on cross-sensory perception with modern instruction tuning for multimodal models and shows that many popular evaluation metrics are insufficient for visually sensory inference.

Diffusion models sample effectively from high-dimensional, multimodal distributions, but their outputs may violate deployment constraints. For task-space robot policies, generated grasps, waypoints, or trajectories can be distributionally valid yet infeasible, violating reachability, collision-avoidance, or closed-loop executability requirements. This embodiment gap limits zero-shot deployment across robots, even when the task-space behavior itself is transferable. We propose an inference-time optimization framework that couples the behavior generation to physical feasibility by formulating diffusion guidance as a constrained optimization problem. Our key insight is to replace the sampling perturbation in the backward process with an optimized correction, allowing hard constraints or soft penalties to be imposed during sampling without the need to retrain the diffusion model, while keeping samples close to the learned prior. We evaluate the method on dexterous grasp synthesis with reachability and collision-avoidance constraints, and dynamic manipulation with controller-level trackability constraints. Across settings and robot embodiments, optimization-guided denoising matches the feasibility of projection- and gradient-guidance baselines while better preserving grasp quality, and improving controller-level executability and task success, with task success improving by up to 20pp. on dexterous grasping and 23pp. on visuomotor manipulation over the best baseline.

We study the problem of recovering the parameters of a multivariate Ornstein-Uhlenbeck (OU) process from steady-state observational and interventional data. In many applications, such as large-scale gene perturbation experiments, only stationary "snapshot" measurements are available, making standard stochastic differential equation estimation methods that rely on time-series trajectories inapplicable. We first establish an identifiability result: one intervention per strongly connected component (SCC) of the drift graph suffices to recover all OU process parameters generically up to a global scaling factor. This holds provided that the SCC condensation graph is connected with a single root and certain spectral nondegeneracy assumptions hold. We propose a recursive learning algorithm that orders SCCs topologically and, for each component, isolates its marginal dynamics and solves a linear system derived from the steady-state moment equations, leveraging parameters recovered for upstream components. Building on this theoretical foundation, we propose a regularized least-squares estimator that jointly minimizes residuals of the steady-state mean and covariance equations across observational and interventional data. Experimental results validate our theoretical findings in recovering parameters of the underlying OU process.

Machine learning-based intrusion detection systems (IDS) are critical for securing Industrial Internet of Things (IIoT) environments. Most adversarial research against them perturbs the feature vector or the traffic that produces it, and depends on gradient access, repeated model queries, or a learned model of benign traffic. A smaller line of work reshapes packet timing without querying the detector, but makes malicious traffic mimic a learned model of benign timing. Across these approaches, one assumption of industrial monitoring pipelines has received little attention: temporal synchronization. An IDS reconstructs operational state by aggregating telemetry into sliding or tumbling windows, so its view depends not only on what is observed but on when each observation falls relative to a window boundary. We introduce the Phantom State Attack (PSA), which exploits that dependence under a passive, zero-query threat model. Rather than modifying packets, perturbing features, querying the classifier, or fitting any model of benign traffic, PSA injects bounded timing drift calibrated to the attack flow's own inter-arrival variability, moving observations across the nearest window boundary by the minimal shift needed. The IDS then reconstructs a phantom state that diverges from the true process state. We evaluate PSA on ToN-IoT and CIC IIoT 2025 (DataSense), against Random Forest, MLP and XGBoost, measuring detection degradation, synchronization distortion, stealth, and attacker cost. PSA degrades detection on flows carrying enough packets for window-boundary redistribution, and leaves others almost unchanged, so its effect is conditional. A query-based baseline reaches higher raw success but needs many queries per window, while PSA needs none. The results identify temporal aggregation as an attack surface reachable under weaker assumptions than prior evasion techniques.

Circuit cutting runs a quantum computation on processors smaller than the circuit, at the price of classical knitting whose cost grows exponentially with the number of cut gates. Keeping this cost affordable is necessary but not sufficient for an advantage, since knitting classically easy subcircuits is itself classically easy. We formulate three requirements for distributed quantum advantage under a stated budget: affordable knitting, classical difficulty that survives the cut, and, for variational circuits, resolvable gradients. We prove that under stated assumptions knitting stays affordable when subcircuits grow behind a bounded interface and becomes exponentially expensive when fixed-width subcircuits multiply. We turn the requirements into screening criteria and apply them to eighteen circuit families. We identify finite local-depth circuits with bounded interfaces as a setting in which the requirements can coexist, with hardness established only in the worst case. As a classically verifiable proof of principle, we knit one gate joining two field-perturbed toric-code patches on an IBM Nighthawk processor for up to 142 spins, and confirm that the reconstruction can at least resolve the contact correlation beyond independent execution for up to 98 qubits.

Wave emissions at frequencies near electron gyrofrequency harmonics are observed at small heliocentric distances below about 40 solar radii and are known to occur in regions with quiescent magnetic fields. We show the close connection of these waves with the large-scale properties of the magnetic field. Near electron gyrofrequency harmonics emissions occur only when the ambient magnetic field points to a narrow range of directions bounded by polar and azimuthal angular ranges in the RTN coordinate system of correspondingly 80^{circ} lesssim θ_B lesssim 100^{circ} and 10^{circ} lesssim φ_B lesssim 30^{circ}. We show that the amplitudes of wave emissions are highest when both angles are close to the center of their respective angular interval favorable to wave emissions. The intensity of wave emissions correlates with the magnetic field angular changes at both large and small time scales. Wave emissions intervals correlate with intervals of decreases in the amplitudes of broadband magnetic fluctuations at low frequencies of 10Hz-100Hz. We discuss possible generation mechanisms of the waves.

Huge corpora of textual data are always known to be a crucial need for training deep models such as transformer-based ones. This issue is emerging more in lower resource languages - like Farsi. We propose naab, the biggest cleaned and ready-to-use open-source textual corpus in Farsi. It contains about 130GB of data, 250 million paragraphs, and 15 billion words. The project name is derived from the Farsi word NAAB K which means pure and high grade. We also provide the raw version of the corpus called naab-raw and an easy-to-use preprocessor that can be employed by those who wanted to make a customized corpus.

Modern image captioning system relies heavily on extracting knowledge from images to capture the concept of a static story. In this paper, we propose a textual visual context dataset for captioning, in which the publicly available dataset COCO Captions (Lin et al., 2014) has been extended with information about the scene (such as objects in the image). Since this information has a textual form, it can be used to leverage any NLP task, such as text similarity or semantic relation methods, into captioning systems, either as an end-to-end training strategy or a post-processing based approach.

Language is inherent and compulsory for human communication. Whether expressed in a written or spoken way, it ensures understanding between people of the same and different regions. With the growing awareness and effort to include more low-resourced languages in NLP research, African languages have recently been a major subject of research in machine translation, and other text-based areas of NLP. However, there is still very little comparable research in speech recognition for African languages. Interestingly, some of the unique properties of African languages affecting NLP, like their diacritical and tonal complexities, have a major root in their speech, suggesting that careful speech interpretation could provide more intuition on how to deal with the linguistic complexities of African languages for text-based NLP. OkwuGb\'e is a step towards building speech recognition systems for African low-resourced languages. Using Fon and Igbo as our case study, we conduct a comprehensive linguistic analysis of each language and describe the creation of end-to-end, deep neural network-based speech recognition models for both languages. We present a state-of-art ASR model for Fon, as well as benchmark ASR model results for Igbo. Our linguistic analyses (for Fon and Igbo) provide valuable insights and guidance into the creation of speech recognition models for other African low-resourced languages, as well as guide future NLP research for Fon and Igbo. The Fon and Igbo models source code have been made publicly available.

We investigate the performance of a state-of-the art (SoTA) architecture T5 (available on the SuperGLUE) and compare with it 3 other previous SoTA architectures across 5 different tasks from 2 relatively diverse datasets. The datasets are diverse in terms of the number and types of tasks they have. To improve performance, we augment the training data by using an autoregressive model. We achieve near-SoTA results on a couple of the tasks - macro F1 scores of 81.66% for task A of the OLID 2019 dataset and 82.54% for task A of the hate speech and offensive content (HASOC) 2021 dataset, where SoTA are 82.9% and 83.05%, respectively. We perform error analysis and explain why one of the models (Bi-LSTM) makes the predictions it does by using a publicly available algorithm: Integrated Gradient (IG). This is because explainable artificial intelligence (XAI) is essential for earning the trust of users. The main contributions of this work are the implementation method of T5, which is discussed; the data augmentation using a new conversational AI model checkpoint, which brought performance improvements; and the revelation on the shortcomings of HASOC 2021 dataset. It reveals the difficulties of poor data annotation by using a small set of examples where the T5 model made the correct predictions, even when the ground truth of the test set were incorrect (in our opinion). We also provide our model checkpoints on the HuggingFace hub1 to foster transparency.

A key trait of daily conversations between individuals is the ability to express empathy towards others, and exploring ways to implement empathy is a crucial step towards human-like dialogue systems. Previous approaches on this topic mainly focus on detecting and utilizing the user's emotion for generating empathetic responses. However, since empathy includes both aspects of affection and cognition, we argue that in addition to identifying the user's emotion, cognitive understanding of the user's situation should also be considered. To this end, we propose a novel approach for empathetic response generation, which leverages commonsense to draw more information about the user's situation and uses this additional information to further enhance the empathy expression in generated responses. We evaluate our approach on EmpatheticDialogues, which is a widely-used benchmark dataset for empathetic response generation. Empirical results demonstrate that our approach outperforms the baseline models in both automatic and human evaluations and can generate more informative and empathetic responses.

Transfer Learning (TL) is an efficient machine learning paradigm that allows overcoming some of the hurdles that characterize the successful training of deep neural networks, ranging from long training times to the needs of large datasets. While exploiting TL is a well established and successful training practice in Supervised Learning (SL), its applicability in Deep Reinforcement Learning (DRL) is rarer. In this paper, we study the level of transferability of three different variants of Deep-Q Networks on popular DRL benchmarks as well as on a set of novel, carefully designed control tasks. Our results show that transferring neural networks in a DRL context can be particularly challenging and is a process which in most cases results in negative transfer. In the attempt of understanding why Deep-Q Networks transfer so poorly, we gain novel insights into the training dynamics that characterizes this family of algorithms.

A capsule is a group of neurons whose activity vector represents the instantiation parameters of a specific type of entity such as an object or an object part. We use the length of the activity vector to represent the probability that the entity exists and its orientation to represent the instantiation parameters. Active capsules at one level make predictions, via transformation matrices, for the instantiation parameters of higher-level capsules. When multiple predictions agree, a higher level capsule becomes active. We show that a discrimininatively trained, multi-layer capsule system achieves state-of-the-art performance on MNIST and is considerably better than a convolutional net at recognizing highly overlapping digits. To achieve these results we use an iterative routing-by-agreement mechanism: A lower-level capsule prefers to send its output to higher level capsules whose activity vectors have a big scalar product with the prediction coming from the lower-level capsule.

In light of recent advancements in generative AI models, it has become essential to distinguish genuine content from AI-generated one to prevent the malicious usage of fake materials as authentic ones and vice versa. Various techniques have been introduced for identifying AI-generated images, with watermarking emerging as a promising approach. In this paper, we analyze the robustness of various AI-image detectors including watermarking and classifier-based deepfake detectors. For watermarking methods that introduce subtle image perturbations (i.e., low perturbation budget methods), we reveal a fundamental trade-off between the evasion error rate (i.e., the fraction of watermarked images detected as non-watermarked ones) and the spoofing error rate (i.e., the fraction of non-watermarked images detected as watermarked ones) upon an application of a diffusion purification attack. In this regime, we also empirically show that diffusion purification effectively removes watermarks with minimal changes to images. For high perturbation watermarking methods where notable changes are applied to images, the diffusion purification attack is not effective. In this case, we develop a model substitution adversarial attack that can successfully remove watermarks. Moreover, we show that watermarking methods are vulnerable to spoofing attacks where the attacker aims to have real images (potentially obscene) identified as watermarked ones, damaging the reputation of the developers. In particular, by just having black-box access to the watermarking method, we show that one can generate a watermarked noise image which can be added to the real images to have them falsely flagged as watermarked ones. Finally, we extend our theory to characterize a fundamental trade-off between the robustness and reliability of classifier-based deep fake detectors and demonstrate it through experiments.

Recent parameter-efficient finetuning (PEFT) techniques aim to improve over the considerable cost of fully finetuning large pretrained language models (PLM). As different PEFT techniques proliferate, it is becoming difficult to compare them, in particular in terms of (i) the structure and functionality they add to the PLM, (ii) the different types and degrees of efficiency improvements achieved, (iii) performance at different downstream tasks, and (iv) how differences in structure and functionality relate to efficiency and task performance. To facilitate such comparisons, this paper presents a reference architecture which standardises aspects shared by different PEFT techniques, while isolating differences to specific locations and interactions with the standard components. Through this process of standardising and isolating differences, a modular view of PEFT techniques emerges, supporting not only direct comparison of different techniques and their efficiency and task performance, but also systematic exploration of reusability and composability of the different types of finetuned modules. We demonstrate how the reference architecture can be applied to understand properties and relative advantages of PEFT techniques, hence to inform selection of techniques for specific tasks, and design choices for new PEFT techniques.

Pre-trained Programming Language Models (PPLMs) achieved many recent states of the art results for many code-related software engineering tasks. Though some studies use data flow or propose tree-based models that utilize Abstract Syntax Tree (AST), most PPLMs do not fully utilize the rich syntactical information in source code. Still, the input is considered a sequence of tokens. There are two issues; the first is computational inefficiency due to the quadratic relationship between input length and attention complexity. Second, any syntactical information, when needed as an extra input to the current PPLMs, requires the model to be pre-trained from scratch, wasting all the computational resources already used for pre-training the current models. In this work, we propose Named Entity Recognition (NER) adapters, lightweight modules that can be inserted into Transformer blocks to learn type information extracted from the AST. These adapters can be used with current PPLMs such as CodeBERT, GraphCodeBERT, and CodeT5. We train the NER adapters using a novel Token Type Classification objective function (TTC). We insert our proposed work in CodeBERT, building CodeBERTER, and evaluate the performance on two tasks of code refinement and code summarization. CodeBERTER improves the accuracy of code refinement from 16.4 to 17.8 while using 20% of training parameter budget compared to the fully fine-tuning approach, and the BLEU score of code summarization from 14.75 to 15.90 while reducing 77% of training parameters compared to the fully fine-tuning approach.

Bone degradation, especially for astronauts in microgravity conditions, is crucial for space exploration missions since the lower applied external forces accelerate the diminution in bone stiffness and strength substantially. Although existing computational models help us understand this phenomenon and possibly restrict its effect in the future, they are time-consuming to simulate the changes in the bones, not just the bone microstructures, of each individual in detail. In this study, a robust yet fast computational method to predict and visualize bone degradation has been developed. Our deep-learning method, TransVNet, can take in different 3D voxelized images and predict their evolution throughout months utilizing a hybrid 3D-CNN-VisionTransformer autoencoder architecture. Because of limited available experimental data and challenges of obtaining new samples, a digital twin dataset of diverse and initial bone-like microstructures was generated to train our TransVNet on the evolution of the 3D images through a previously developed degradation model for microgravity.

Recent advances in Large Language Models (LLMs) have highlighted the need for robust, comprehensive, and challenging benchmarks. Yet, research on evaluating their Emotional Intelligence (EI) is considerably limited. Existing benchmarks have two major shortcomings: first, they mainly focus on emotion recognition, neglecting essential EI capabilities such as emotion regulation and thought facilitation through emotion understanding; second, they are primarily constructed from existing datasets, which include frequent patterns, explicit information, and annotation errors, leading to unreliable evaluation. We propose EmoBench, a benchmark that draws upon established psychological theories and proposes a comprehensive definition for machine EI, including Emotional Understanding and Emotional Application. EmoBench includes a set of 400 hand-crafted questions in English and Chinese, which are meticulously designed to require thorough reasoning and understanding. Our findings reveal a considerable gap between the EI of existing LLMs and the average human, highlighting a promising direction for future research. Our code and data will be publicly available from https://github.com/Sahandfer/EmoBench.

We present a method to build animatable dog avatars from monocular videos. This is challenging as animals display a range of (unpredictable) non-rigid movements and have a variety of appearance details (e.g., fur, spots, tails). We develop an approach that links the video frames via a 4D solution that jointly solves for animal's pose variation, and its appearance (in a canonical pose). To this end, we significantly improve the quality of template-based shape fitting by endowing the SMAL parametric model with Continuous Surface Embeddings, which brings image-to-mesh reprojection constaints that are denser, and thus stronger, than the previously used sparse semantic keypoint correspondences. To model appearance, we propose an implicit duplex-mesh texture that is defined in the canonical pose, but can be deformed using SMAL pose coefficients and later rendered to enforce a photometric compatibility with the input video frames. On the challenging CoP3D and APTv2 datasets, we demonstrate superior results (both in terms of pose estimates and predicted appearance) to existing template-free (RAC) and template-based approaches (BARC, BITE).

Prediction of stock prices plays a significant role in aiding the decision-making of investors. Considering its importance, a growing literature has emerged trying to forecast stock prices with improved accuracy. In this study, we introduce an innovative approach for forecasting stock prices with greater accuracy. We incorporate external economic environment-related information along with stock prices. In our novel approach, we improve the performance of stock price prediction by taking into account variations due to future expected macroeconomic policy changes as investors adjust their current behavior ahead of time based on expected future macroeconomic policy changes. Furthermore, we incorporate macroeconomic variables along with historical stock prices to make predictions. Results from this strongly support the inclusion of future economic policy changes along with current macroeconomic information. We confirm the supremacy of our method over the conventional approach using several tree-based machine-learning algorithms. Results are strongly conclusive across various machine learning models. Our preferred model outperforms the conventional approach with an RMSE value of 1.61 compared to an RMSE value of 1.75 from the conventional approach.

Diffusion models (DMs) have established themselves as the state-of-the-art generative modeling approach in the visual domain and beyond. A crucial drawback of DMs is their slow sampling speed, relying on many sequential function evaluations through large neural networks. Sampling from DMs can be seen as solving a differential equation through a discretized set of noise levels known as the sampling schedule. While past works primarily focused on deriving efficient solvers, little attention has been given to finding optimal sampling schedules, and the entire literature relies on hand-crafted heuristics. In this work, for the first time, we propose a general and principled approach to optimizing the sampling schedules of DMs for high-quality outputs, called Align Your Steps. We leverage methods from stochastic calculus and find optimal schedules specific to different solvers, trained DMs and datasets. We evaluate our novel approach on several image, video as well as 2D toy data synthesis benchmarks, using a variety of different samplers, and observe that our optimized schedules outperform previous hand-crafted schedules in almost all experiments. Our method demonstrates the untapped potential of sampling schedule optimization, especially in the few-step synthesis regime.

The recent advances in deep-learning have led to the development of highly sophisticated systems with an unquenchable appetite for data. On the other hand, building good deep-learning models for low-resource languages remains a challenging task. This paper focuses on developing a Question Answering dataset for two such languages- Hindi and Marathi. Despite Hindi being the 3rd most spoken language worldwide, with 345 million speakers, and Marathi being the 11th most spoken language globally, with 83.2 million speakers, both languages face limited resources for building efficient Question Answering systems. To tackle the challenge of data scarcity, we have developed a novel approach for translating the SQuAD 2.0 dataset into Hindi and Marathi. We release the largest Question-Answering dataset available for these languages, with each dataset containing 28,000 samples. We evaluate the dataset on various architectures and release the best-performing models for both Hindi and Marathi, which will facilitate further research in these languages. Leveraging similarity tools, our method holds the potential to create datasets in diverse languages, thereby enhancing the understanding of natural language across varied linguistic contexts. Our fine-tuned models, code, and dataset will be made publicly available.

In this paper, we investigate the impact of objects on gender bias in image captioning systems. Our results show that only gender-specific objects have a strong gender bias (e.g., women-lipstick). In addition, we propose a visual semantic-based gender score that measures the degree of bias and can be used as a plug-in for any image captioning system. Our experiments demonstrate the utility of the gender score, since we observe that our score can measure the bias relation between a caption and its related gender; therefore, our score can be used as an additional metric to the existing Object Gender Co-Occ approach. Code and data are publicly available at https://github.com/ahmedssabir/GenderScore.

In this work, we focus on improving the captions generated by image-caption generation systems. We propose a novel re-ranking approach that leverages visual-semantic measures to identify the ideal caption that maximally captures the visual information in the image. Our re-ranker utilizes the Belief Revision framework (Blok et al., 2003) to calibrate the original likelihood of the top-n captions by explicitly exploiting the semantic relatedness between the depicted caption and the visual context. Our experiments demonstrate the utility of our approach, where we observe that our re-ranker can enhance the performance of a typical image-captioning system without the necessity of any additional training or fine-tuning.

3D Gaussian Splatting (3DGS) is a promising technique for 3D reconstruction, offering efficient training and rendering speeds, making it suitable for real-time applications.However, current methods require highly controlled environments (no moving people or wind-blown elements, and consistent lighting) to meet the inter-view consistency assumption of 3DGS. This makes reconstruction of real-world captures problematic. We present SpotlessSplats, an approach that leverages pre-trained and general-purpose features coupled with robust optimization to effectively ignore transient distractors. Our method achieves state-of-the-art reconstruction quality both visually and quantitatively, on casual captures.

Accurate prediction of protein-ligand binding affinities is crucial in drug discovery, particularly during hit-to-lead and lead optimization phases, however, limitations in ligand force fields continue to impact prediction accuracy. In this work, we validate relative binding free energy (RBFE) accuracy using neural network potentials (NNPs) for the ligands. We utilize a novel NNP model, AceForce 1.0, based on the TensorNet architecture for small molecules that broadens the applicability to diverse drug-like compounds, including all important chemical elements and supporting charged molecules. Using established benchmarks, we show overall improved accuracy and correlation in binding affinity predictions compared with GAFF2 for molecular mechanics and ANI2-x for NNPs. Slightly less accuracy but comparable correlations with OPLS4. We also show that we can run the NNP simulations at 2 fs timestep, at least two times larger than previous NNP models, providing significant speed gains. The results show promise for further evolutions of free energy calculations using NNPs while demonstrating its practical use already with the current generation. The code and NNP model are publicly available for research use.

Artificial Intelligence (AI) systems are increasingly intertwined with daily life, assisting users in executing various tasks and providing guidance on decision-making. This integration introduces risks of AI-driven manipulation, where such systems may exploit users' cognitive biases and emotional vulnerabilities to steer them toward harmful outcomes. Through a randomized controlled trial with 233 participants, we examined human susceptibility to such manipulation in financial (e.g., purchases) and emotional (e.g., conflict resolution) decision-making contexts. Participants interacted with one of three AI agents: a neutral agent (NA) optimizing for user benefit without explicit influence, a manipulative agent (MA) designed to covertly influence beliefs and behaviors, or a strategy-enhanced manipulative agent (SEMA) employing explicit psychological tactics to reach its hidden objectives. By analyzing participants' decision patterns and shifts in their preference ratings post-interaction, we found significant susceptibility to AI-driven manipulation. Particularly, across both decision-making domains, participants interacting with the manipulative agents shifted toward harmful options at substantially higher rates (financial, MA: 62.3%, SEMA: 59.6%; emotional, MA: 42.3%, SEMA: 41.5%) compared to the NA group (financial, 35.8%; emotional, 12.8%). Notably, our findings reveal that even subtle manipulative objectives (MA) can be as effective as employing explicit psychological strategies (SEMA) in swaying human decision-making. By revealing the potential for covert AI influence, this study highlights a critical vulnerability in human-AI interactions, emphasizing the need for ethical safeguards and regulatory frameworks to ensure responsible deployment of AI technologies and protect human autonomy.

The growing demand for mental health support has highlighted the importance of conversational agents as human supporters worldwide and in China. These agents could increase availability and reduce the relative costs of mental health support. The provided support can be divided into two main types: cognitive and emotional support. Existing work on this topic mainly focuses on constructing agents that adopt Cognitive Behavioral Therapy (CBT) principles. Such agents operate based on pre-defined templates and exercises to provide cognitive support. However, research on emotional support using such agents is limited. In addition, most of the constructed agents operate in English, highlighting the importance of conducting such studies in China. In this study, we analyze the effectiveness of Emohaa in reducing symptoms of mental distress. Emohaa is a conversational agent that provides cognitive support through CBT-based exercises and guided conversations. It also emotionally supports users by enabling them to vent their desired emotional problems. The study included 134 participants, split into three groups: Emohaa (CBT-based), Emohaa (Full), and control. Experimental results demonstrated that compared to the control group, participants who used Emohaa experienced considerably more significant improvements in symptoms of mental distress. We also found that adding the emotional support agent had a complementary effect on such improvements, mainly depression and insomnia. Based on the obtained results and participants' satisfaction with the platform, we concluded that Emohaa is a practical and effective tool for reducing mental distress.

Diffusion- and flow-based models have emerged as state-of-the-art generative modeling approaches, but they require many sampling steps. Consistency models can distill these models into efficient one-step generators; however, unlike flow- and diffusion-based methods, their performance inevitably degrades when increasing the number of steps, which we show both analytically and empirically. Flow maps generalize these approaches by connecting any two noise levels in a single step and remain effective across all step counts. In this paper, we introduce two new continuous-time objectives for training flow maps, along with additional novel training techniques, generalizing existing consistency and flow matching objectives. We further demonstrate that autoguidance can improve performance, using a low-quality model for guidance during distillation, and an additional boost can be achieved by adversarial finetuning, with minimal loss in sample diversity. We extensively validate our flow map models, called Align Your Flow, on challenging image generation benchmarks and achieve state-of-the-art few-step generation performance on both ImageNet 64x64 and 512x512, using small and efficient neural networks. Finally, we show text-to-image flow map models that outperform all existing non-adversarially trained few-step samplers in text-conditioned synthesis.

Reconstructing dynamic assets from video data is central to many in computer vision and graphics tasks. Existing 4D reconstruction approaches are limited by category-specific models or slow optimization-based methods. Inspired by the recent Large Reconstruction Model (LRM), we present the Large Interpolation Model (LIM), a transformer-based feed-forward solution, guided by a novel causal consistency loss, for interpolating implicit 3D representations across time. Given implicit 3D representations at times t_0 and t_1, LIM produces a deformed shape at any continuous time tin[t_0,t_1], delivering high-quality interpolated frames in seconds. Furthermore, LIM allows explicit mesh tracking across time, producing a consistently uv-textured mesh sequence ready for integration into existing production pipelines. We also use LIM, in conjunction with a diffusion-based multiview generator, to produce dynamic 4D reconstructions from monocular videos. We evaluate LIM on various dynamic datasets, benchmarking against image-space interpolation methods (e.g., FiLM) and direct triplane linear interpolation, and demonstrate clear advantages. In summary, LIM is the first feed-forward model capable of high-speed tracked 4D asset reconstruction across diverse categories.

In this work, we investigate the correlation between gender and contextual biases, focusing on elements such as action verbs, object nouns, and particularly on occupations. We introduce a novel dataset, GenderLexicon, and a framework that can estimate contextual bias and its related gender bias. Our model can interpret the bias with a score and thus improve the explainability of gender bias. Also, our findings confirm the existence of gender biases beyond occupational stereotypes. To validate our approach and demonstrate its effectiveness, we conduct evaluations on five diverse datasets, including a Japanese dataset.

As more speech technologies rely on a supervised deep learning approach with clean speech as the ground truth, a methodology to onboard said speech at scale is needed. However, this approach needs to minimize the dependency on human listening and annotation, only requiring a human-in-the-loop when needed. In this paper, we address this issue by outlining Speech Enhancement-based Curation Pipeline (SECP) which serves as a framework to onboard clean speech. This clean speech can then train a speech enhancement model, which can further refine the original dataset and thus close the iterative loop. By running two iterative rounds, we observe that enhanced output used as ground truth does not degrade model performance according to Delta_{PESQ}, a metric used in this paper. We also show through comparative mean opinion score (CMOS) based subjective tests that the highest and lowest bound of refined data is perceptually better than the original data.

This study presents a quantitative evaluation of the code quality and security of five prominent Large Language Models (LLMs): Claude Sonnet 4, Claude 3.7 Sonnet, GPT-4o, Llama 3.2 90B, and OpenCoder 8B. While prior research has assessed the functional performance of LLM-generated code, this research tested LLM output from 4,442 Java coding assignments through comprehensive static analysis using SonarQube. The findings suggest that although LLMs can generate functional code, they also introduce a range of software defects, including bugs, security vulnerabilities, and code smells. These defects do not appear to be isolated; rather, they may represent shared weaknesses stemming from systemic limitations within current LLM code generation methods. In particular, critically severe issues, such as hard-coded passwords and path traversal vulnerabilities, were observed across multiple models. These results indicate that LLM-generated code requires verification in order to be considered production-ready. This study found no direct correlation between a model's functional performance (measured by Pass@1 rate of unit tests) and the overall quality and security of its generated code, measured by the number of SonarQube issues in benchmark solutions that passed the functional tests. This suggests that functional benchmark performance score is not a good indicator of overall code quality and security. The goal of this study is not to rank LLM performance but to highlight that all evaluated models appear to share certain weaknesses. Consequently, these findings support the view that static analysis can be a valuable instrument for detecting latent defects and an important safeguard for organizations that deploy AI in software development.

Efficient deployment of deep learning models for aerial object detection on resource-constrained devices requires significant compression without com-promising performance. In this study, we propose a novel three-stage compression pipeline for the YOLOv8 object detection model, integrating sparsity-aware training, structured channel pruning, and Channel-Wise Knowledge Distillation (CWD). First, sparsity-aware training introduces dynamic sparsity during model optimization, effectively balancing parameter reduction and detection accuracy. Second, we apply structured channel pruning by leveraging batch normalization scaling factors to eliminate redundant channels, significantly reducing model size and computational complexity. Finally, to mitigate the accuracy drop caused by pruning, we employ CWD to transfer knowledge from the original model, using an adjustable temperature and loss weighting scheme tailored for small and medium object detection. Extensive experiments on the VisDrone dataset demonstrate the effectiveness of our approach across multiple YOLOv8 variants. For YOLOv8m, our method reduces model parameters from 25.85M to 6.85M (a 73.51% reduction), FLOPs from 49.6G to 13.3G, and MACs from 101G to 34.5G, while reducing AP50 by only 2.7%. The resulting compressed model achieves 47.9 AP50 and boosts inference speed from 26 FPS (YOLOv8m baseline) to 45 FPS, enabling real-time deployment on edge devices. We further apply TensorRT as a lightweight optimization step. While this introduces a minor drop in AP50 (from 47.9 to 47.6), it significantly improves inference speed from 45 to 68 FPS, demonstrating the practicality of our approach for high-throughput, re-source-constrained scenarios.

A common recipe to improve diffusion models at test-time so that samples score highly against a user-specified reward is to introduce the gradient of the reward into the dynamics of the diffusion itself. This procedure is often ill posed, as user-specified rewards are usually only well defined on the data distribution at the end of generation. While common workarounds to this problem are to use a denoiser to estimate what a sample would have been at the end of generation, we propose a simple solution to this problem by working directly with a flow map. By exploiting a relationship between the flow map and velocity field governing the instantaneous transport, we construct an algorithm, Flow Map Trajectory Tilting (FMTT), which provably performs better ascent on the reward than standard test-time methods involving the gradient of the reward. The approach can be used to either perform exact sampling via importance weighting or principled search that identifies local maximizers of the reward-tilted distribution. We demonstrate the efficacy of our approach against other look-ahead techniques, and show how the flow map enables engagement with complicated reward functions that make possible new forms of image editing, e.g. by interfacing with vision language models.

Generating animated 3D objects is at the heart of many applications, yet most advanced works are typically difficult to apply in practice because of their limited setup, their long runtime, or their limited quality. We introduce ActionMesh, a generative model that predicts production-ready 3D meshes "in action" in a feed-forward manner. Drawing inspiration from early video models, our key insight is to modify existing 3D diffusion models to include a temporal axis, resulting in a framework we dubbed "temporal 3D diffusion". Specifically, we first adapt the 3D diffusion stage to generate a sequence of synchronized latents representing time-varying and independent 3D shapes. Second, we design a temporal 3D autoencoder that translates a sequence of independent shapes into the corresponding deformations of a pre-defined reference shape, allowing us to build an animation. Combining these two components, ActionMesh generates animated 3D meshes from different inputs like a monocular video, a text description, or even a 3D mesh with a text prompt describing its animation. Besides, compared to previous approaches, our method is fast and produces results that are rig-free and topology consistent, hence enabling rapid iteration and seamless applications like texturing and retargeting. We evaluate our model on standard video-to-4D benchmarks (Consistent4D, Objaverse) and report state-of-the-art performances on both geometric accuracy and temporal consistency, demonstrating that our model can deliver animated 3D meshes with unprecedented speed and quality.

The increased use of Large Language Models (LLMs) in sensitive domains leads to growing interest in how their confidence scores correspond to fairness and bias. This study examines the alignment between LLM-predicted confidence and human-annotated bias judgments. Focusing on gender bias, the research investigates probability confidence calibration in contexts involving gendered pronoun resolution. The goal is to evaluate if calibration metrics based on predicted confidence scores effectively capture fairness-related disparities in LLMs. The results show that, among the six state-of-the-art models, Gemma-2 demonstrates the worst calibration according to the gender bias benchmark. The primary contribution of this work is a fairness-aware evaluation of LLMs' confidence calibration, offering guidance for ethical deployment. In addition, we introduce a new calibration metric, Gender-ECE, designed to measure gender disparities in resolution tasks.

In this paper, we examine the thermodynamic behavior of a quantum harmonic oscillator with a position-dependent mass (PDM), where spatial inhomogeneity is modeled through a deformation parameter α. Based on the exact energy spectrum, we explore the resulting thermodynamic quantities and superstatistics. Our findings reveal that increasing α leads to a decrease in entropy and specific heat, reflecting a confinement-induced reduction in the number of accessible states. The partition function and free energy exhibit smooth behavior across all parameter regimes, indicating the absence of critical phase transitions. This study underscores the influence of mass deformation on quantum thermal responses and demonstrates that, while the overall thermodynamic trends are consistent with those reported in the literature, certain distinctive features emerge due to the specific form of the deformation.

Social interactions are fundamental to well-being, yet automatically detecting them in daily life-particularly using wearables-remains underexplored. Most existing systems are evaluated in controlled settings, focus primarily on in-person interactions, or rely on restrictive assumptions (e.g., requiring multiple speakers within fixed temporal windows), limiting generalizability to real-world use. We present an on-watch interaction detection system designed to capture diverse interactions in naturalistic settings. A core component is a foreground speech detector trained on a public dataset. Evaluated on over 100,000 labeled foreground speech and background sound instances, the detector achieves a balanced accuracy of 85.51%, outperforming prior work by 5.11%. We evaluated the system in a real-world deployment (N=38), with over 900 hours of total smartwatch wear time. The system detected 1,691 interactions, 77.28% were confirmed via participant self-report, with durations ranging from under one minute to over one hour. Among correct detections, 81.45% were in-person, 15.7% virtual, and 1.85% hybrid. Leveraging participant-labeled data, we further developed a multimodal model achieving a balanced accuracy of 90.36% and a sensitivity of 91.17% on 33,698 labeled 15-second windows. These results demonstrate the feasibility of real-world interaction sensing and open the door to adaptive, context-aware systems responding to users' dynamic social environments.

Large Language Models (LLMs) and Code-LLMs (CLLMs) have significantly improved code generation, but, they frequently face difficulties when dealing with challenging and complex problems. Retrieval-Augmented Generation (RAG) addresses this issue by retrieving and integrating external knowledge at the inference time. However, retrieval models often fail to find most relevant context, and generation models, with limited context capacity, can hallucinate when given irrelevant data. We present a novel framework that leverages a Programming Knowledge Graph (PKG) to semantically represent and retrieve code. This approach enables fine-grained code retrieval by focusing on the most relevant segments while reducing irrelevant context through a tree-pruning technique. PKG is coupled with a re-ranking mechanism to reduce even more hallucinations by selectively integrating non-RAG solutions. We propose two retrieval approaches-block-wise and function-wise-based on the PKG, optimizing context granularity. Evaluations on the HumanEval and MBPP benchmarks show our method improves pass@1 accuracy by up to 20%, and outperforms state-of-the-art models by up to 34% on MBPP. Our contributions include PKG-based retrieval, tree pruning to enhance retrieval precision, a re-ranking method for robust solution selection and a Fill-in-the-Middle (FIM) enhancer module for automatic code augmentation with relevant comments and docstrings.

Automatically generating source code from natural language using large language models (LLMs) is becoming common, yet security vulnerabilities persist despite advances in fine tuning and prompting. In this work, we systematically evaluate whether multi LLM ensembles and collaborative strategies can meaningfully improve secure code generation. We present MULTI-LLMSECCODEEVAL, a framework for assessing and enhancing security across the vulnerability management lifecycle by combining multiple LLMs with static analysis and structured collaboration. Using SecLLMEval and SecLLMHolmes, we benchmark ten pipelines spanning single model, ensemble, collaborative, and hybrid designs. Our results show that ensemble pipelines augmented with static analysis improve secure code generation over single LLM baselines by up to 47.3% on SecLLMEval and 19.3% on SecLLMHolmes, while purely LLM based collaborative pipelines yield smaller gains of 8.9% to 22.3%. Hybrid pipelines that integrate ensembling, detection, and patching achieve the strongest security performance, outperforming the best ensemble baseline by 1.78% to 4.72% and collaborative baselines by 19.81% to 26.78%. Ablation studies reveal that model scale alone does not ensure security. Smaller, structured multi model ensembles consistently outperform large monolithic LLMs. Overall, our findings demonstrate that secure code does not emerge from scale, but from carefully orchestrated multi model system design.

Mechanism-targeted synthetic data is increasingly proposed as a way to steer pretraining toward desirable capabilities, but it remains unclear how such interventions should be evaluated. We study this question for in-context learning (ICL) under matched compute (iso-FLOPs) using Bi-Induct, a lightweight data rewrite that interleaves short directional copy snippets into a natural pretraining stream: forward-copy (induction), backward-copy (anti-induction, as a directional control), or a balanced mix. Across 0.13B-1B decoder-only models, we evaluate (i) few-shot performance on standard LM benchmarks and function-style ICL probes, (ii) head-level copy telemetry, and (iii) held-out perplexity as a guardrail. Bi-Induct reliably increases induction-head activity, but this does not translate into consistent improvements in few-shot generalization: on standard LM benchmarks, Bi-Induct is largely performance-neutral relative to natural-only training, while on function-style probes the 1B natural-only model performs best. Despite explicit backward-copy cues, anti-induction scores remain near zero across scales, revealing a strong forward/backward asymmetry. Targeted ablations show a sharper distinction: removing the top 2% induction heads per layer harms ICL more than matched random ablations, with the largest relative drop occurring in the natural-only models. This indicates that natural-only training produces more centralized, load-bearing induction circuitry, whereas Bi-Induct tends to create more distributed and redundant induction activity. Our main conclusion is that eliciting a mechanism is not the same as making it load-bearing. For data-centric foundation model design, this suggests that synthetic data interventions should be evaluated not only by signature amplification, but by whether they create causally necessary computation while preserving natural-data modeling quality.

Large language models are commonly trained through multi-stage post-training: first via RLHF, then fine-tuned for other downstream objectives. Yet even small downstream updates can compromise earlier learned behaviors (e.g., safety), exposing a brittleness known as catastrophic forgetting. This suggests standard RLHF objectives do not guarantee robustness to future adaptation. To address it, most prior work designs downstream-time methods to preserve previously learned behaviors. We argue that preventing this requires pre-finetuning robustness: the base policy should avoid brittle high-reward solutions whose reward drops sharply under standard fine-tuning. We propose Fine-tuning Robust Policy Optimization (FRPO), a robust RLHF framework that optimizes reward not only at the current policy, but across a KL-bounded neighborhood of policies reachable by downstream adaptation. The key idea is to ensure reward stability under policy shifts via a max-min formulation. By modifying GRPO, we develop an algorithm with no extra computation, and empirically show it substantially reduces safety degradation across multiple base models and downstream fine-tuning regimes (SFT and RL) while preserving downstream task performance. We further study a math-focused RL setting, demonstrating that FRPO preserves accuracy under subsequent fine-tuning.

Supervised fine-tuning (SFT) provides the standard approach for teaching LLMs new behaviors from offline expert demonstrations. However, standard SFT uniformly fits all samples -- including those with low likelihood under the base model -- which can disproportionately drive training updates toward overfitting specific samples rather than learning the target behavior. Moreover, adapting to these unlikely samples induces substantial policy shifts that degrade prior capabilities. Existing methods mitigate this by filtering, regenerating, or down-weighting low-likelihood data. In doing so, they often suppress precisely the novel behaviors the base model has yet to learn. We propose InfoSFT, a principled weighting scheme for the SFT objective that concentrates learning signals on maximally informative, medium-confidence tokens -- those neither overly familiar to the base model nor too unlikely to cause instability. Requiring only a one-line modification to the standard token-wise loss, InfoSFT demonstrably improves generalization over vanilla SFT and likelihood-weighted baselines across math, code, and chain-of-thought tasks with diverse model families, while better preserving pre-existing capabilities.

We want to detect whether a particular image dataset has been used to train a model. We propose a new technique, radioactive data, that makes imperceptible changes to this dataset such that any model trained on it will bear an identifiable mark. The mark is robust to strong variations such as different architectures or optimization methods. Given a trained model, our technique detects the use of radioactive data and provides a level of confidence (p-value). Our experiments on large-scale benchmarks (Imagenet), using standard architectures (Resnet-18, VGG-16, Densenet-121) and training procedures, show that we can detect usage of radioactive data with high confidence (p<10^-4) even when only 1% of the data used to trained our model is radioactive. Our method is robust to data augmentation and the stochasticity of deep network optimization. As a result, it offers a much higher signal-to-noise ratio than data poisoning and backdoor methods.

High-performance GPU kernels are critical to modern machine learning systems, yet developing them remains a manual, expert-driven process. Recent work has explored using LLMs to automate kernel generation, but generated kernels still fall short of carefully tuned references on standardized benchmarks. We present CuTeGen, an agentic GPU kernel synthesis framework that treats kernel development as a structured generate-test-refine workflow over the CuTe abstraction layer. Two design choices distinguish CuTeGen from prior work: targeting CuTe rather than raw CUDA, which exposes performance-critical structures such as tiling and data movement while remaining stable enough for iterative refinement, and a delayed profiling schedule that withholds low-level performance feedback until the kernel's high-level structure has stabilized. On the 209 tasks of KernelBench Level-1 and Level-2, CuTeGen achieves an average speedup of 1.71times over PyTorch and outperforms the prior agentic baseline CudaForge (0.89times) at comparable per-task generation cost. Code available at https://github.com/taratt/cutegen.git

Recent computer-using-agent (CUA) red-teaming papers report prompt-injection attack success rates (ASR) of 42-98%, but these headline numbers cluster on retired models and on the most-vulnerable model in each paper's panel. We ask whether those techniques, reproduced as hand-crafted templates, still work against current frontier CUAs. We release CUA-HandCrafted, a public benchmark of 793 episodes spanning 24 multi-step web tasks, 56 attack templates, 8 attack families, and 4 system-prompt configurations. Against Claude Sonnet 4.6 and GPT-5.4 we measure 0/140 multi-step attack success (Clopper-Pearson 95% upper bound 2.60%); a prompt ablation shows this resistance lives in the model weights. Yet it does not generalize: on a sister coding-agent benchmark (SkillBench), the same weights fall to hand-crafted skill-injection at up to 100%. We argue that the literature's high ASR is largely attributable to RL-optimized injection text rather than the attack categories, and that frontier safety hardening is domain-conditioned, specific to the heavily-targeted browser surface. Reporting techniques without releasing the optimized strings, or extrapolating browser-domain safety to other CUA modalities, makes published ASR numbers unreproducible.

LLM-based simulated clients are increasingly used to train novice counselors, evaluate LLM therapists, and generate synthetic data. However, current simulators produce overly cooperative clients that disclose too readily, accept therapeutic reframes without resistance, and resolve core issues within a single session. We trace these issues to profiles that lack causal depth and behavioral mechanisms that treat all content as equally accessible. We present PatientAct, a framework for client simulation grounded in established clinical theories. Our profiles integrate the 5Ps clinical case formulation, providing causal depth without tying the design to any single therapeutic modality. During simulation, profiles include a dynamic memory layer in which items carry trust thresholds (e.g., symptoms are available early, whereas formative memories require a sustained therapeutic alliance). At each turn, the client's emotional reaction and behavior are modeled before generating a response. If the therapist approaches gated content, PatientAct expresses resistance in terms of quantity, content, and style rather than defaulting to cooperation or a single resistance pattern. We evaluate our framework on 40 clinical situations and demonstrate that it generates diverse profiles with high clinical plausibility. Moreover, PatientAct significantly outperforms the baselines, yielding substantial gains in resistance quality and behavioral realism. Our code and data will be publicly available via github.com/Sahandfer/PatientHub.

MRI reconstruction from undersampled k-space measurements is an ill-posed inverse problem. Physics-driven deep learning (PD-DL) methods have shown strong performance for this task by combining the MRI forward model with learned image regularization within algorithm-unrolling frameworks. However, most existing PD-DL methods reconstruct complex-valued images directly, thereby implicitly coupling magnitude and phase within a single learned representation. This coupled regularization may be suboptimal in reconstruction settings where accurate phase modeling plays an important role, such as partial Fourier (PF) imaging, where recovery of the omitted asymmetric k-space measurements depends on the underlying image phase. In such scenarios, explicit modeling of magnitude and phase as separate components may reduce the reliance on externally estimated or predefined phase information. To this end, we propose UMPIRE-Net (Unrolled Magnitude-Phase In REgularization Network), a PD-DL method that introduces separate learned regularizers for magnitude and phase components, together with a novel data-fidelity formulation that enforces measurements consistency. We evaluate UMPIRE-Net for accelerated MRI with PF across different datasets and acceleration factors. Experimental results demonstrate that our proposed method improves reconstruction quality compared with a conventional complex-valued PD-DL baseline, yielding sharper images and reduced artifacts. Code available at: https://github.com/MahdiSaberii/UMPIRE-Net

Structure-from-Motion (SfM) is a cornerstone of 3D perception, yet current methods often fail when applied to complex videos involving challenging camera motions or dynamic scenes. Compounding the problem, the field lacks reliable ground-truth benchmarks for such difficult scenarios, making it hard to gauge real-world progress or to pinpoint where improvements are most needed. To address this gap, we introduce a new benchmark for evaluating camera pose estimation. Our key insight is to leverage online panoramic 360° video as a source of data from which to construct challenging clips, while still enabling robust ground-truth trajectory recovery. The panoramic nature of these videos provides richer visual context for tracking camera motion, even when parts of the view are affected by blur, motion, or dynamic objects. After tracking camera motion across full 360° videos, we crop and reproject selected portions to generate perspective-view clips that serve as our benchmark, called ORBIT. Experiments show that COLMAP, as well as recent optimization-based and feed-forward SfM methods struggle to accurately estimate camera poses on our benchmark. Hence, ORBIT provides a valuable testbed where researchers can meaningfully measure progress on truly challenging, real-world SfM problems.

Recent advances in large language model (LLM) pretraining highlight the role of high-quality training data in improving performance. While model-based filtering has proven effective in selecting high-quality subsets from web-scale corpora, especially for high-resource languages, low-resource languages face challenges due to limited availability of annotated data. This work explores extending quality filtering to over 100 languages by proposing a multilingual adaptation approach that converts an existing English quality classifier into a multilingual variant. Our approach proposes training a small multi-layer perceptron on top of Transformer encoder-only model embeddings, using multilingual text as input and scores obtained from English classifiers applied to machine-translated text as labels. Our 1B, 3B and 8B scale experiments show that our approach maintains the downstream LLM benchmark performance of existing multilingual model-based filtering baselines, without harming regional and cultural knowledge benchmarks. To further evaluate cross-lingual generalization, we compare classifier scores of high-quality synthetic data and web samples, and the correlation of classifier scores with LLM-based ones, revealing that the classifier can learn the scoring criteria of its original English variant, even for languages not included in its training data.

Text-based world models have recently shown potential to simulate complex RL environment state transitions for a diverse set of use cases. However, no standardized benchmark exists to compare, study and improve these models. Through this paper, we attempt to solve this problem by building GlobeBench, a novel benchmark with 1,632 tasks across five separate domains and nineteen simulation skills. The tasks check whether models can predict the environment's response to a specific action given the full agent trajectory by understanding earlier state changes, following environment rules, and predicting the effects of actions on both related and unrelated objects. We compare the predicted environment response with an execution-derived reference, using a hybrid architecture with both deterministic gates and a semantic judge. Among 20 general-purpose LLMs and specialized text world models, GPT-6 Astra achieves the highest task success rate at 49.57%. We find that model accuracy decreases as actions affect a larger portion of the environment. Specifically, when the agent's changes must propagate to dependent objects, task success drops to an average of 5.1% across all 20 models. Although model strengths differ across tasks, 631 tasks remain unsolved by any model, suggesting a significant gap in general LLM capabilities as a world model.

In this paper, we find a sample complexity bound for learning a simplex from noisy samples. Assume a dataset of size n is given which includes i.i.d. samples drawn from a uniform distribution over an unknown simplex in R^K, where samples are assumed to be corrupted by a multi-variate additive Gaussian noise of an arbitrary magnitude. We prove the existence of an algorithm that with high probability outputs a simplex having a ell_2 distance of at most varepsilon from the true simplex (for any varepsilon>0). Also, we theoretically show that in order to achieve this bound, it is sufficient to have ngeleft(K^2/varepsilon^2right)e^{Omegaleft(K/SNR^2right)} samples, where SNR stands for the signal-to-noise ratio. This result solves an important open problem and shows as long as SNRgeOmegaleft(K^{1/2}right), the sample complexity of the noisy regime has the same order to that of the noiseless case. Our proofs are a combination of the so-called sample compression technique in ashtiani2018nearly, mathematical tools from high-dimensional geometry, and Fourier analysis. In particular, we have proposed a general Fourier-based technique for recovery of a more general class of distribution families from additive Gaussian noise, which can be further used in a variety of other related problems.

Learning neural implicit surfaces from volume rendering has become popular for multi-view reconstruction. Neural surface reconstruction approaches can recover complex 3D geometry that are difficult for classical Multi-view Stereo (MVS) approaches, such as non-Lambertian surfaces and thin structures. However, one key assumption for these methods is knowing accurate camera parameters for the input multi-view images, which are not always available. In this paper, we present NoPose-NeuS, a neural implicit surface reconstruction method that extends NeuS to jointly optimize camera poses with the geometry and color networks. We encode the camera poses as a multi-layer perceptron (MLP) and introduce two additional losses, which are multi-view feature consistency and rendered depth losses, to constrain the learned geometry for better estimated camera poses and scene surfaces. Extensive experiments on the DTU dataset show that the proposed method can estimate relatively accurate camera poses, while maintaining a high surface reconstruction quality with 0.89 mean Chamfer distance.

Word alignments are useful for tasks like statistical and neural machine translation (NMT) and cross-lingual annotation projection. Statistical word aligners perform well, as do methods that extract alignments jointly with translations in NMT. However, most approaches require parallel training data, and quality decreases as less training data is available. We propose word alignment methods that require no parallel data. The key idea is to leverage multilingual word embeddings, both static and contextualized, for word alignment. Our multilingual embeddings are created from monolingual data only without relying on any parallel data or dictionaries. We find that alignments created from embeddings are superior for four and comparable for two language pairs compared to those produced by traditional statistical aligners, even with abundant parallel data; e.g., contextualized embeddings achieve a word alignment F1 for English-German that is 5 percentage points higher than eflomal, a high-quality statistical aligner, trained on 100k parallel sentences.

Recently, there has been a growing interest in utilizing machine learning for accurate classification of power quality events (PQEs). However, most of these studies are performed assuming an ideal situation, while in reality, we can have measurement noise, DC offset, and variations in the voltage signal's amplitude and frequency. Building on the prior PQE classification works using deep learning, this paper proposes a deep-learning framework that leverages attention-enabled Transformers as a tool to accurately classify PQEs under the aforementioned considerations. The proposed framework can operate directly on the voltage signals with no need for a separate feature extraction or calculation phase. Our results show that the proposed framework outperforms recently proposed learning-based techniques. It can accurately classify PQEs under the aforementioned conditions with an accuracy varying between 99.81%-91.43% depending on the signal-to-noise ratio, DC offsets, and variations in the signal amplitude and frequency.

We propose a novel framework for incorporating unlabeled data into semi-supervised classification problems, where scenarios involving the minimization of either i) adversarially robust or ii) non-robust loss functions have been considered. Notably, we allow the unlabeled samples to deviate slightly (in total variation sense) from the in-domain distribution. The core idea behind our framework is to combine Distributionally Robust Optimization (DRO) with self-supervised training. As a result, we also leverage efficient polynomial-time algorithms for the training stage. From a theoretical standpoint, we apply our framework on the classification problem of a mixture of two Gaussians in R^d, where in addition to the m independent and labeled samples from the true distribution, a set of n (usually with ngg m) out of domain and unlabeled samples are given as well. Using only the labeled data, it is known that the generalization error can be bounded by proptoleft(d/mright)^{1/2}. However, using our method on both isotropic and non-isotropic Gaussian mixture models, one can derive a new set of analytically explicit and non-asymptotic bounds which show substantial improvement on the generalization error compared to ERM. Our results underscore two significant insights: 1) out-of-domain samples, even when unlabeled, can be harnessed to narrow the generalization gap, provided that the true data distribution adheres to a form of the ``cluster assumption", and 2) the semi-supervised learning paradigm can be regarded as a special case of our framework when there are no distributional shifts. We validate our claims through experiments conducted on a variety of synthetic and real-world datasets.

We deal with the combinatorial problem of learning directed acyclic graph (DAG) structure from observational data adhering to a linear structural equation model (SEM). Leveraging advances in differentiable, nonconvex characterizations of acyclicity, recent efforts have advocated a continuous constrained optimization paradigm to efficiently explore the space of DAGs. Most existing methods employ lasso-type score functions to guide this search, which (i) require expensive penalty parameter retuning when the unknown SEM noise variances change across problem instances; and (ii) implicitly rely on limiting homoscedasticity assumptions. In this work, we propose a new convex score function for sparsity-aware learning of linear DAGs, which incorporates concomitant estimation of scale and thus effectively decouples the sparsity parameter from the exogenous noise levels. Regularization via a smooth, nonconvex acyclicity penalty term yields CoLiDE (Concomitant Linear DAG Estimation), a regression-based criterion amenable to efficient gradient computation and closed-form estimation of noise variances in heteroscedastic scenarios. Our algorithm outperforms state-of-the-art methods without incurring added complexity, especially when the DAGs are larger and the noise level profile is heterogeneous. We also find CoLiDE exhibits enhanced stability manifested via reduced standard deviations in several domain-specific metrics, underscoring the robustness of our novel linear DAG estimator.

Human-Robot Interaction (HRI) has become increasingly important as robots are being integrated into various aspects of daily life. One key aspect of HRI is gesture recognition, which allows robots to interpret and respond to human gestures in real-time. Gesture recognition plays an important role in non-verbal communication in HRI. To this aim, there is ongoing research on how such non-verbal communication can strengthen verbal communication and improve the system's overall efficiency, thereby enhancing the user experience with the robot. However, several challenges need to be addressed in gesture recognition systems, which include data generation, transferability, scalability, generalizability, standardization, and lack of benchmarking of the gestural systems. In this preliminary paper, we want to address the challenges of data generation using virtual reality simulations and standardization issues by presenting gestures to some commands that can be used as a standard in ground robots.

Large Language Models (LLMs) have demonstrated unprecedented capabilities across various natural language processing tasks. Their ability to process and generate viable text and code has made them ubiquitous in many fields, while their deployment as knowledge bases and "reasoning" tools remains an area of ongoing research. In geography, a growing body of literature has been focusing on evaluating LLMs' geographical knowledge and their ability to perform spatial reasoning. However, very little is still known about the internal functioning of these models, especially about how they process geographical information. In this chapter, we establish a novel framework for the study of geospatial mechanistic interpretability - using spatial analysis to reverse engineer how LLMs handle geographical information. Our aim is to advance our understanding of the internal representations that these complex models generate while processing geographical information - what one might call "how LLMs think about geographic information" if such phrasing was not an undue anthropomorphism. We first outline the use of probing in revealing internal structures within LLMs. We then introduce the field of mechanistic interpretability, discussing the superposition hypothesis and the role of sparse autoencoders in disentangling polysemantic internal representations of LLMs into more interpretable, monosemantic features. In our experiments, we use spatial autocorrelation to show how features obtained for placenames display spatial patterns related to their geographic location and can thus be interpreted geospatially, providing insights into how these models process geographical information. We conclude by discussing how our framework can help shape the study and use of foundation models in geography.

The task of style transfer for 3D Gaussian splats has been explored in many previous works, but these require reconstructing or fine-tuning the splat while incorporating style information or optimizing a feature extraction network on the splat representation. We propose a reconstruction- and optimization-free approach to stylizing 3D Gaussian splats. This is done by generating a graph structure across the implicit surface of the splat representation. A feed-forward, surface-based stylization method is then used and interpolated back to the individual splats in the scene. This allows for any style image and 3D Gaussian splat to be used without any additional training or optimization. This also allows for fast stylization of splats, achieving speeds under 2 minutes even on consumer-grade hardware. We demonstrate the quality results this approach achieves and compare to other 3D Gaussian splat style transfer methods. Code is publicly available at https://github.com/davidmhart/FastSplatStyler.

We introduce a family of random matrices where correlations between matrix elements are induced via interaction-derived Boltzmann factors. Varying these yields access to different ensembles. We find a universal scaling behavior of the finite-size statistics characterized by a heavy-tailed eigenvalue distribution whose extremes are governed by the Fréchet extreme value distribution for the case corresponding to a ferromagnetic Ising transition. The introduction of a finite density of nonlocal interactions restores standard random-matrix behavior. Suitably rescaled average extremes, playing a physical role as an order parameter, can thus discriminate aspects of the interaction structure; they also yield further nonuniversal information. In particular, the link between maximum eigenvalues and order parameters offers a potential route to resolving long-standing problems in statistical physics, such as deriving the exact magnetization scaling function in the two-dimensional Ising model at criticality.

Pretraining Bidirectional Encoder Representations from Transformers (BERT) for downstream NLP tasks is a non-trival task. We pretrained 5 BERT models that differ in the size of their training sets, mixture of formal and informal Arabic, and linguistic preprocessing. All are intended to support Arabic dialects and social media. The experiments highlight the centrality of data diversity and the efficacy of linguistically aware segmentation. They also highlight that more data or more training step do not necessitate better models. Our new models achieve new state-of-the-art results on several downstream tasks. The resulting models are released to the community under the name QARiB.

We revisit watermarking techniques based on pre-trained deep networks, in the light of self-supervised approaches. We present a way to embed both marks and binary messages into their latent spaces, leveraging data augmentation at marking time. Our method can operate at any resolution and creates watermarks robust to a broad range of transformations (rotations, crops, JPEG, contrast, etc). It significantly outperforms the previous zero-bit methods, and its performance on multi-bit watermarking is on par with state-of-the-art encoder-decoder architectures trained end-to-end for watermarking. The code is available at github.com/facebookresearch/ssl_watermarking

Crowd-sourcing is commonly adopted for dialog data collection. However, it is highly costly and time-consuming, and the collected data is limited in scale and topic coverage. In this paper, aiming to generate emotional support conversations, we propose exploiting large-scale pre-trained language models for data augmentation, and provide key findings in our pilot exploration. Our adopted approach leverages the 6B-parameter GPT-J model and utilizes publicly available dialog posts to trigger conversations on various topics. Then we construct AugESC, a machine-augmented dataset for emotional support conversation. It is two orders of magnitude larger than the original ESConv dataset in scale, covers more diverse topics, and is shown to be of high quality by human evaluation. Lastly, we demonstrate with interactive evaluation that AugESC can further enhance dialog models tuned on ESConv to handle various conversation topics and to provide significantly more effective emotional support.

Differential privacy (DP) is by far the most widely accepted framework for mitigating privacy risks in machine learning. However, exactly how small the privacy parameter epsilon needs to be to protect against certain privacy risks in practice is still not well-understood. In this work, we study data reconstruction attacks for discrete data and analyze it under the framework of multiple hypothesis testing. We utilize different variants of the celebrated Fano's inequality to derive upper bounds on the inferential power of a data reconstruction adversary when the model is trained differentially privately. Importantly, we show that if the underlying private data takes values from a set of size M, then the target privacy parameter epsilon can be O(log M) before the adversary gains significant inferential power. Our analysis offers theoretical evidence for the empirical effectiveness of DP against data reconstruction attacks even at relatively large values of epsilon.

With Artificial Intelligence (AI) increasingly permeating various aspects of society, including healthcare, the adoption of the Transformers neural network architecture is rapidly changing many applications. Transformer is a type of deep learning architecture initially developed to solve general-purpose Natural Language Processing (NLP) tasks and has subsequently been adapted in many fields, including healthcare. In this survey paper, we provide an overview of how this architecture has been adopted to analyze various forms of data, including medical imaging, structured and unstructured Electronic Health Records (EHR), social media, physiological signals, and biomolecular sequences. Those models could help in clinical diagnosis, report generation, data reconstruction, and drug/protein synthesis. We identified relevant studies using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. We also discuss the benefits and limitations of using transformers in healthcare and examine issues such as computational cost, model interpretability, fairness, alignment with human values, ethical implications, and environmental impact.

Due to the lack of human resources for mental health support, there is an increasing demand for employing conversational agents for support. Recent work has demonstrated the effectiveness of dialogue models in providing emotional support. As previous studies have demonstrated that seekers' persona is an important factor for effective support, we investigate whether there are benefits to modeling such information in dialogue models for support. In this paper, our empirical analysis verifies that persona has an important impact on emotional support. Therefore, we propose a framework for dynamically inferring and modeling seekers' persona. We first train a model for inferring the seeker's persona from the conversation history. Accordingly, we propose PAL, a model that leverages persona information and, in conjunction with our strategy-based controllable generation method, provides personalized emotional support. Automatic and manual evaluations demonstrate that PAL achieves state-of-the-art results, outperforming the baselines on the studied benchmark. Our code and data are publicly available at https://github.com/chengjl19/PAL.

Most state-of-the-art instance segmentation methods rely on large amounts of pixel-precise ground-truth annotations for training, which are expensive to create. Interactive segmentation networks help generate such annotations based on an image and the corresponding user interactions such as clicks. Existing methods for this task can only process a single instance at a time and each user interaction requires a full forward pass through the entire deep network. We introduce a more efficient approach, called DynaMITe, in which we represent user interactions as spatio-temporal queries to a Transformer decoder with a potential to segment multiple object instances in a single iteration. Our architecture also alleviates any need to re-compute image features during refinement, and requires fewer interactions for segmenting multiple instances in a single image when compared to other methods. DynaMITe achieves state-of-the-art results on multiple existing interactive segmentation benchmarks, and also on the new multi-instance benchmark that we propose in this paper.

We introduce the first multitasking vision transformer adapters that learn generalizable task affinities which can be applied to novel tasks and domains. Integrated into an off-the-shelf vision transformer backbone, our adapters can simultaneously solve multiple dense vision tasks in a parameter-efficient manner, unlike existing multitasking transformers that are parametrically expensive. In contrast to concurrent methods, we do not require retraining or fine-tuning whenever a new task or domain is added. We introduce a task-adapted attention mechanism within our adapter framework that combines gradient-based task similarities with attention-based ones. The learned task affinities generalize to the following settings: zero-shot task transfer, unsupervised domain adaptation, and generalization without fine-tuning to novel domains. We demonstrate that our approach outperforms not only the existing convolutional neural network-based multitasking methods but also the vision transformer-based ones. Our project page is at https://ivrl.github.io/VTAGML.

Distinct-n scoreLi2016 is a widely used automatic metric for evaluating diversity in language generation tasks. However, we observed that the original approach for calculating distinct scores has evident biases that tend to assign higher penalties to longer sequences. We refine the calculation of distinct scores by scaling the number of distinct tokens based on their expectations. We provide both empirical and theoretical evidence to show that our method effectively removes the biases existing in the original distinct score. Our experiments show that our proposed metric, Expectation-Adjusted Distinct (EAD), correlates better with human judgment in evaluating response diversity. To foster future research, we provide an example implementation at https://github.com/lsy641/Expectation-Adjusted-Distinct.

A plethora of text-guided image editing methods have recently been developed by leveraging the impressive capabilities of large-scale diffusion-based generative models such as Imagen and Stable Diffusion. A standardized evaluation protocol, however, does not exist to compare methods across different types of fine-grained edits. To address this gap, we introduce EditVal, a standardized benchmark for quantitatively evaluating text-guided image editing methods. EditVal consists of a curated dataset of images, a set of editable attributes for each image drawn from 13 possible edit types, and an automated evaluation pipeline that uses pre-trained vision-language models to assess the fidelity of generated images for each edit type. We use EditVal to benchmark 8 cutting-edge diffusion-based editing methods including SINE, Imagic and Instruct-Pix2Pix. We complement this with a large-scale human study where we show that EditVall's automated evaluation pipeline is strongly correlated with human-preferences for the edit types we considered. From both the human study and automated evaluation, we find that: (i) Instruct-Pix2Pix, Null-Text and SINE are the top-performing methods averaged across different edit types, however {\it only} Instruct-Pix2Pix and Null-Text are able to preserve original image properties; (ii) Most of the editing methods fail at edits involving spatial operations (e.g., changing the position of an object). (iii) There is no `winner' method which ranks the best individually across a range of different edit types. We hope that our benchmark can pave the way to developing more reliable text-guided image editing tools in the future. We will publicly release EditVal, and all associated code and human-study templates to support these research directions in https://deep-ml-research.github.io/editval/.

Morality is a fundamental aspect of human behavior and ethics, influencing how we interact with each other and the world around us. When faced with a moral dilemma, a person's ability to make clear moral judgments can be clouded. Due to many factors such as personal biases, emotions and situational factors people can find it difficult to decide their best course of action. The AmITheAsshole (AITA) subreddit is a forum on the social media platform Reddit that helps people get clarity and objectivity on their predicaments. In the forum people post anecdotes about moral dilemmas they are facing in their lives, seeking validation for their actions or advice on how to navigate the situation from the community. The morality of the actions in each post is classified based on the collective opinion of the community into mainly two labels, "Not The Asshole" (NTA) and "You Are The Asshole" (YTA). This project aims to generate comments with moral reasoning for stories with moral dilemmas using the AITA subreddit as a dataset. While past literature has explored the classification of posts into labels (Alhassan et al., 2022), the generation of comments remains a novel and challenging task. It involves understanding the complex social and ethical considerations in each situation. To address this challenge, we will leverage the vast amount of data on the forum with the goal of generating coherent comments that align with the norms and values of the AITA community. In this endeavor, we aim to evaluate state-of-the-art seq2seq text generation models for their ability to make moral judgments similarly to humans, ultimately producing concise comments providing clear moral stances and advice for the poster.