id
string
title
string
abstract
string
text
string
label
string
primary_category
string
published
string
2511.05301v2
QueStER: Query Specification for Generative keyword-based Retrieval
Generative retrieval (GR) differs from the traditional index-then-retrieve pipeline by storing relevance in model parameters and generating retrieval cues directly from the query, but it can be brittle out of domain and expensive to scale. We introduce QueStER (QUEry SpecificaTion for gEnerative Keyword-Based Retrieval...
QueStER: Query Specification for Generative keyword-based Retrieval. Generative retrieval (GR) differs from the traditional index-then-retrieve pipeline by storing relevance in model parameters and generating retrieval cues directly from the query, but it can be brittle out of domain and expensive to scale. We introduc...
ir_retrieval
cs.IR
2025-11-07T15:01:38+00:00
2601.00075v1
IMBWatch -- a Spatio-Temporal Graph Neural Network approach to detect Illicit Massage Business
Illicit Massage Businesses (IMBs) are a covert and persistent form of organized exploitation that operate under the facade of legitimate wellness services while facilitating human trafficking, sexual exploitation, and coerced labor. Detecting IMBs is difficult due to encoded digital advertisements, frequent changes in ...
IMBWatch -- a Spatio-Temporal Graph Neural Network approach to detect Illicit Massage Business. Illicit Massage Businesses (IMBs) are a covert and persistent form of organized exploitation that operate under the facade of legitimate wellness services while facilitating human trafficking, sexual exploitation, and coerce...
lg_ml
cs.LG
2025-12-31T19:06:41+00:00
2512.19243v3
VisionDirector: Vision-Language Guided Closed-Loop Refinement for Generative Image Synthesis
Generative models can now produce photorealistic imagery, yet they still struggle with the long, multi-goal prompts that professional designers issue. To expose this gap and better evaluate models' performance in real-world settings, we introduce Long Goal Bench (LGBench), a 2,000-task suite (1,000 T2I and 1,000 I2I) w...
VisionDirector: Vision-Language Guided Closed-Loop Refinement for Generative Image Synthesis. Generative models can now produce photorealistic imagery, yet they still struggle with the long, multi-goal prompts that professional designers issue. To expose this gap and better evaluate models' performance in real-world se...
cv_vision
cs.CV
2025-12-22T10:25:38+00:00
2512.23983v1
DriveExplorer: Images-Only Decoupled 4D Reconstruction with Progressive Restoration for Driving View Extrapolation
This paper presents an effective solution for view extrapolation in autonomous driving scenarios. Recent approaches focus on generating shifted novel view images from given viewpoints using diffusion models. However, these methods heavily rely on priors such as LiDAR point clouds, 3D bounding boxes, and lane annotation...
DriveExplorer: Images-Only Decoupled 4D Reconstruction with Progressive Restoration for Driving View Extrapolation. This paper presents an effective solution for view extrapolation in autonomous driving scenarios. Recent approaches focus on generating shifted novel view images from given viewpoints using diffusion mode...
cv_vision
cs.CV
2025-12-30T04:41:56+00:00
2512.16105v1
BayesSum: Bayesian Quadrature in Discrete Spaces
This paper addresses the challenging computational problem of estimating intractable expectations over discrete domains. Existing approaches, including Monte Carlo and Russian Roulette estimators, are consistent but often require a large number of samples to achieve accurate results. We propose a novel estimator, \emph...
BayesSum: Bayesian Quadrature in Discrete Spaces. This paper addresses the challenging computational problem of estimating intractable expectations over discrete domains. Existing approaches, including Monte Carlo and Russian Roulette estimators, are consistent but often require a large number of samples to achieve acc...
stat_ml
stat.ML
2025-12-18T02:43:30+00:00
2511.14096v1
NeuroPath: Neurobiology-Inspired Path Tracking and Reflection for Semantically Coherent Retrieval
Retrieval-augmented generation (RAG) greatly enhances large language models (LLMs) performance in knowledge-intensive tasks. However, naive RAG methods struggle with multi-hop question answering due to their limited capacity to capture complex dependencies across documents. Recent studies employ graph-based RAG to capt...
NeuroPath: Neurobiology-Inspired Path Tracking and Reflection for Semantically Coherent Retrieval. Retrieval-augmented generation (RAG) greatly enhances large language models (LLMs) performance in knowledge-intensive tasks. However, naive RAG methods struggle with multi-hop question answering due to their limited capac...
ir_retrieval
cs.IR
2025-11-18T03:28:23+00:00
2512.17570v2
GreedySnake: Accelerating SSD-Offloaded LLM Training with Efficient Scheduling and Optimizer Step Overlapping
SSD-offloaded training offers a practical and promising approach to making LLM training cost-effective. Building on gradient accumulation with micro-batches, this paper introduces GreedySnake, a new SSD-offloaded training system that employs vertical scheduling, which executes all microbatches of a layer before proceed...
GreedySnake: Accelerating SSD-Offloaded LLM Training with Efficient Scheduling and Optimizer Step Overlapping. SSD-offloaded training offers a practical and promising approach to making LLM training cost-effective. Building on gradient accumulation with micro-batches, this paper introduces GreedySnake, a new SSD-offloa...
lg_ml
cs.LG
2025-12-19T13:36:31+00:00
2510.08335v2
PAC Learnability in the Presence of Performativity
Following the wide-spread adoption of machine learning models in real-world applications, the phenomenon of performativity, i.e. model-dependent shifts in the test distribution, becomes increasingly prevalent. Unfortunately, since models are usually trained solely based on samples from the original (unshifted) distribu...
PAC Learnability in the Presence of Performativity. Following the wide-spread adoption of machine learning models in real-world applications, the phenomenon of performativity, i.e. model-dependent shifts in the test distribution, becomes increasingly prevalent. Unfortunately, since models are usually trained solely bas...
stat_ml
stat.ML
2025-10-09T15:22:52+00:00
2512.24517v2
Paragraph Segmentation Revisited: Towards a Standard Task for Structuring Speech
Automatic speech transcripts are often delivered as unstructured word streams that impede readability and repurposing. We recast paragraph segmentation as the missing structuring step and fill three gaps at the intersection of speech processing and text segmentation. First, we establish TEDPara (human-annotated TED tal...
Paragraph Segmentation Revisited: Towards a Standard Task for Structuring Speech. Automatic speech transcripts are often delivered as unstructured word streams that impede readability and repurposing. We recast paragraph segmentation as the missing structuring step and fill three gaps at the intersection of speech proc...
cl_nlp
cs.CL
2025-12-30T23:29:51+00:00
2512.24551v3
PhyGDPO: Physics-Aware Groupwise Direct Preference Optimization for Physically Consistent Text-to-Video Generation
Recent advances in text-to-video (T2V) generation have achieved good visual quality, yet synthesizing videos that faithfully follow physical laws remains an open challenge. Existing methods mainly based on graphics or prompt extension struggle to generalize beyond simple simulated environments or learn implicit physica...
PhyGDPO: Physics-Aware Groupwise Direct Preference Optimization for Physically Consistent Text-to-Video Generation. Recent advances in text-to-video (T2V) generation have achieved good visual quality, yet synthesizing videos that faithfully follow physical laws remains an open challenge. Existing methods mainly based o...
cv_vision
cs.CV
2025-12-31T01:19:14+00:00
2511.00176v1
Effectiveness of LLMs in Temporal User Profiling for Recommendation
Effectively modeling the dynamic nature of user preferences is crucial for enhancing recommendation accuracy and fostering transparency in recommender systems. Traditional user profiling often overlooks the distinction between transitory short-term interests and stable long-term preferences. This paper examines the cap...
Effectiveness of LLMs in Temporal User Profiling for Recommendation. Effectively modeling the dynamic nature of user preferences is crucial for enhancing recommendation accuracy and fostering transparency in recommender systems. Traditional user profiling often overlooks the distinction between transitory short-term in...
ir_retrieval
cs.IR
2025-10-31T18:28:40+00:00
2601.00848v1
Temporal Attack Pattern Detection in Multi-Agent AI Workflows: An Open Framework for Training Trace-Based Security Models
We present an openly documented methodology for fine-tuning language models to detect temporal attack patterns in multi-agent AI workflows using OpenTelemetry trace analysis. We curate a dataset of 80,851 examples from 18 public cybersecurity sources and 35,026 synthetic OpenTelemetry traces. We apply iterative QLoRA f...
Temporal Attack Pattern Detection in Multi-Agent AI Workflows: An Open Framework for Training Trace-Based Security Models. We present an openly documented methodology for fine-tuning language models to detect temporal attack patterns in multi-agent AI workflows using OpenTelemetry trace analysis. We curate a dataset of...
ai_general
cs.AI
2025-12-29T09:41:22+00:00
2509.02327v3
Variational Uncertainty Decomposition for In-Context Learning
As large language models (LLMs) gain popularity in conducting prediction tasks in-context, understanding the sources of uncertainty in in-context learning becomes essential to ensuring reliability. The recent hypothesis of in-context learning performing predictive Bayesian inference opens the avenue for Bayesian uncert...
Variational Uncertainty Decomposition for In-Context Learning. As large language models (LLMs) gain popularity in conducting prediction tasks in-context, understanding the sources of uncertainty in in-context learning becomes essential to ensuring reliability. The recent hypothesis of in-context learning performing pre...
stat_ml
stat.ML
2025-09-02T13:53:09+00:00
2512.22266v2
LLMTM: Benchmarking and Optimizing LLMs for Temporal Motif Analysis in Dynamic Graphs
The widespread application of Large Language Models (LLMs) has motivated a growing interest in their capacity for processing dynamic graphs. Temporal motifs, as an elementary unit and important local property of dynamic graphs which can directly reflect anomalies and unique phenomena, are essential for understanding th...
LLMTM: Benchmarking and Optimizing LLMs for Temporal Motif Analysis in Dynamic Graphs. The widespread application of Large Language Models (LLMs) has motivated a growing interest in their capacity for processing dynamic graphs. Temporal motifs, as an elementary unit and important local property of dynamic graphs which ...
lg_ml
cs.LG
2025-12-24T18:10:29+00:00
2512.14048v1
Intention Chain-of-Thought Prompting with Dynamic Routing for Code Generation
Large language models (LLMs) exhibit strong generative capabilities and have shown great potential in code generation. Existing chain-of-thought (CoT) prompting methods enhance model reasoning by eliciting intermediate steps, but suffer from two major limitations: First, their uniform application tends to induce overth...
Intention Chain-of-Thought Prompting with Dynamic Routing for Code Generation. Large language models (LLMs) exhibit strong generative capabilities and have shown great potential in code generation. Existing chain-of-thought (CoT) prompting methods enhance model reasoning by eliciting intermediate steps, but suffer from...
ai_general
cs.AI
2025-12-16T03:30:21+00:00
2512.12435v1
Co-Hub Node Based Multiview Graph Learning with Theoretical Guarantees
Identifying the graphical structure underlying the observed multivariate data is essential in numerous applications. Current methodologies are predominantly confined to deducing a singular graph under the presumption that the observed data are uniform. However, many contexts involve heterogeneous datasets that feature ...
Co-Hub Node Based Multiview Graph Learning with Theoretical Guarantees. Identifying the graphical structure underlying the observed multivariate data is essential in numerous applications. Current methodologies are predominantly confined to deducing a singular graph under the presumption that the observed data are unif...
stat_ml
stat.ML
2025-12-13T19:25:35+00:00
2601.03265v1
Jailbreak-Zero: A Path to Pareto Optimal Red Teaming for Large Language Models
This paper introduces Jailbreak-Zero, a novel red teaming methodology that shifts the paradigm of Large Language Model (LLM) safety evaluation from a constrained example-based approach to a more expansive and effective policy-based framework. By leveraging an attack LLM to generate a high volume of diverse adversarial ...
Jailbreak-Zero: A Path to Pareto Optimal Red Teaming for Large Language Models. This paper introduces Jailbreak-Zero, a novel red teaming methodology that shifts the paradigm of Large Language Model (LLM) safety evaluation from a constrained example-based approach to a more expansive and effective policy-based framewor...
cl_nlp
cs.CL
2025-12-18T16:26:09+00:00
2512.25034v1
Generative Classifiers Avoid Shortcut Solutions
Discriminative approaches to classification often learn shortcuts that hold in-distribution but fail even under minor distribution shift. This failure mode stems from an overreliance on features that are spuriously correlated with the label. We show that generative classifiers, which use class-conditional generative mo...
Generative Classifiers Avoid Shortcut Solutions. Discriminative approaches to classification often learn shortcuts that hold in-distribution but fail even under minor distribution shift. This failure mode stems from an overreliance on features that are spuriously correlated with the label. We show that generative class...
lg_ml
cs.LG
2025-12-31T18:31:46+00:00
2510.21021v1
Gaussian Mixture Flow Matching with Domain Alignment for Multi-Domain Sequential Recommendation
Users increasingly interact with content across multiple domains, resulting in sequential behaviors marked by frequent and complex transitions. While Cross-Domain Sequential Recommendation (CDSR) models two-domain interactions, Multi-Domain Sequential Recommendation (MDSR) introduces significantly more domain transitio...
Gaussian Mixture Flow Matching with Domain Alignment for Multi-Domain Sequential Recommendation. Users increasingly interact with content across multiple domains, resulting in sequential behaviors marked by frequent and complex transitions. While Cross-Domain Sequential Recommendation (CDSR) models two-domain interacti...
ir_retrieval
cs.IR
2025-10-23T22:11:26+00:00
2512.24327v2
Topological Spatial Graph Coarsening
Spatial graphs are particular graphs for which the nodes are localized in space (e.g., public transport network, molecules, branching biological structures). In this work, we consider the problem of spatial graph reduction, that aims to find a smaller spatial graph (i.e., with less nodes) with the same overall structur...
Topological Spatial Graph Coarsening. Spatial graphs are particular graphs for which the nodes are localized in space (e.g., public transport network, molecules, branching biological structures). In this work, we consider the problem of spatial graph reduction, that aims to find a smaller spatial graph (i.e., with less...
stat_ml
stat.ML
2025-12-30T16:27:48+00:00
2509.03187v1
Enhancing Interpretability and Effectiveness in Recommendation with Numerical Features via Learning to Contrast the Counterfactual samples
We propose a general model-agnostic Contrastive learning framework with Counterfactual Samples Synthesizing (CCSS) for modeling the monotonicity between the neural network output and numerical features which is critical for interpretability and effectiveness of recommender systems. CCSS models the monotonicity via a tw...
Enhancing Interpretability and Effectiveness in Recommendation with Numerical Features via Learning to Contrast the Counterfactual samples. We propose a general model-agnostic Contrastive learning framework with Counterfactual Samples Synthesizing (CCSS) for modeling the monotonicity between the neural network output a...
ir_retrieval
cs.IR
2025-09-03T10:17:41+00:00
2512.24478v2
HOLOGRAPH: Active Causal Discovery via Sheaf-Theoretic Alignment of Large Language Model Priors
Causal discovery from observational data remains fundamentally limited by identifiability constraints. Recent work has explored leveraging Large Language Models (LLMs) as sources of prior causal knowledge, but existing approaches rely on heuristic integration that lacks theoretical grounding. We introduce HOLOGRAPH, a ...
HOLOGRAPH: Active Causal Discovery via Sheaf-Theoretic Alignment of Large Language Model Priors. Causal discovery from observational data remains fundamentally limited by identifiability constraints. Recent work has explored leveraging Large Language Models (LLMs) as sources of prior causal knowledge, but existing appr...
lg_ml
cs.LG
2025-12-30T21:47:05+00:00
2512.15740v1
The Principle of Proportional Duty: A Knowledge-Duty Framework for Ethical Equilibrium in Human and Artificial Systems
Traditional ethical frameworks often struggle to model decision-making under uncertainty, treating it as a simple constraint on action. This paper introduces the Principle of Proportional Duty (PPD), a novel framework that models how ethical responsibility scales with an agent's epistemic state. The framework reveals t...
The Principle of Proportional Duty: A Knowledge-Duty Framework for Ethical Equilibrium in Human and Artificial Systems. Traditional ethical frameworks often struggle to model decision-making under uncertainty, treating it as a simple constraint on action. This paper introduces the Principle of Proportional Duty (PPD), ...
ai_general
cs.AI
2025-12-07T02:37:07+00:00
2512.04480v7
Prescriptive Artificial Intelligence: A Formal Paradigm for Auditing Human Decisions Under Uncertainty
We formalize Prescriptive Artificial Intelligence as a distinct paradigm for human-AI decision collaboration in high-stakes environments. Unlike predictive systems optimized for outcome accuracy, prescriptive systems are designed to recommend and audit human decisions under uncertainty, providing normative guidance whi...
Prescriptive Artificial Intelligence: A Formal Paradigm for Auditing Human Decisions Under Uncertainty. We formalize Prescriptive Artificial Intelligence as a distinct paradigm for human-AI decision collaboration in high-stakes environments. Unlike predictive systems optimized for outcome accuracy, prescriptive systems...
ai_general
cs.AI
2025-12-04T05:33:28+00:00
2512.21507v5
SVBench: Evaluation of Video Generation Models on Social Reasoning
Recent text-to-video generation models have made remarkable progress in visual realism, motion fidelity, and text-video alignment, yet they still struggle to produce socially coherent behavior. Unlike humans, who readily infer intentions, beliefs, emotions, and social norms from brief visual cues, current models often ...
SVBench: Evaluation of Video Generation Models on Social Reasoning. Recent text-to-video generation models have made remarkable progress in visual realism, motion fidelity, and text-video alignment, yet they still struggle to produce socially coherent behavior. Unlike humans, who readily infer intentions, beliefs, emot...
cv_vision
cs.CV
2025-12-25T04:44:59+00:00
2510.15706v1
GraphMind: Interactive Novelty Assessment System for Accelerating Scientific Discovery
Large Language Models (LLMs) show strong reasoning and text generation capabilities, prompting their use in scientific literature analysis, including novelty assessment. While evaluating novelty of scientific papers is crucial for peer review, it requires extensive knowledge of related work, something not all reviewers...
GraphMind: Interactive Novelty Assessment System for Accelerating Scientific Discovery. Large Language Models (LLMs) show strong reasoning and text generation capabilities, prompting their use in scientific literature analysis, including novelty assessment. While evaluating novelty of scientific papers is crucial for p...
ir_retrieval
cs.IR
2025-10-17T14:49:07+00:00
2512.23707v1
Training AI Co-Scientists Using Rubric Rewards
AI co-scientists are emerging as a tool to assist human researchers in achieving their research goals. A crucial feature of these AI co-scientists is the ability to generate a research plan given a set of aims and constraints. The plan may be used by researchers for brainstorming, or may even be implemented after furth...
Training AI Co-Scientists Using Rubric Rewards. AI co-scientists are emerging as a tool to assist human researchers in achieving their research goals. A crucial feature of these AI co-scientists is the ability to generate a research plan given a set of aims and constraints. The plan may be used by researchers for brain...
lg_ml
cs.LG
2025-12-29T18:59:33+00:00
2510.12054v1
MIARec: Mutual-influence-aware Heterogeneous Network Embedding for Scientific Paper Recommendation
With the rapid expansion of scientific literature, scholars increasingly demand precise and high-quality paper recommendations. Among various recommendation methodologies, graph-based approaches have garnered attention by effectively exploiting the structural characteristics inherent in scholarly networks. However, the...
MIARec: Mutual-influence-aware Heterogeneous Network Embedding for Scientific Paper Recommendation. With the rapid expansion of scientific literature, scholars increasingly demand precise and high-quality paper recommendations. Among various recommendation methodologies, graph-based approaches have garnered attention b...
ir_retrieval
cs.IR
2025-10-14T01:47:25+00:00
2512.15226v1
Yes-MT's Submission to the Low-Resource Indic Language Translation Shared Task in WMT 2024
This paper presents the systems submitted by the Yes-MT team for the Low-Resource Indic Language Translation Shared Task at WMT 2024 (Pakray et al., 2024), focusing on translating between English and the Assamese, Mizo, Khasi, and Manipuri languages. The experiments explored various approaches, including fine-tuning pr...
Yes-MT's Submission to the Low-Resource Indic Language Translation Shared Task in WMT 2024. This paper presents the systems submitted by the Yes-MT team for the Low-Resource Indic Language Translation Shared Task at WMT 2024 (Pakray et al., 2024), focusing on translating between English and the Assamese, Mizo, Khasi, a...
cl_nlp
cs.CL
2025-12-17T09:24:05+00:00
2511.01064v2
Generalized Guarantees for Variational Inference in the Presence of Even and Elliptical Symmetry
We extend several recent results providing symmetry-based guarantees for variational inference (VI) with location-scale families. VI approximates a target density $p$ by the best match $q^*$ in a family $Q$ of tractable distributions that in general does not contain $p$. It is known that VI can recover key properties o...
Generalized Guarantees for Variational Inference in the Presence of Even and Elliptical Symmetry. We extend several recent results providing symmetry-based guarantees for variational inference (VI) with location-scale families. VI approximates a target density $p$ by the best match $q^*$ in a family $Q$ of tractable di...
stat_ml
stat.ML
2025-11-02T20:10:57+00:00
2512.10453v1
Grammaticality Judgments in Humans and Language Models: Revisiting Generative Grammar with LLMs
What counts as evidence for syntactic structure? In traditional generative grammar, systematic contrasts in grammaticality such as subject-auxiliary inversion and the licensing of parasitic gaps are taken as evidence for an internal, hierarchical grammar. In this paper, we test whether large language models (LLMs), tra...
Grammaticality Judgments in Humans and Language Models: Revisiting Generative Grammar with LLMs. What counts as evidence for syntactic structure? In traditional generative grammar, systematic contrasts in grammaticality such as subject-auxiliary inversion and the licensing of parasitic gaps are taken as evidence for an...
cl_nlp
cs.CL
2025-12-11T09:17:35+00:00
2512.22295v1
Learning Dynamic Scene Reconstruction with Sinusoidal Geometric Priors
We propose SirenPose, a novel loss function that combines the periodic activation properties of sinusoidal representation networks with geometric priors derived from keypoint structures to improve the accuracy of dynamic 3D scene reconstruction. Existing approaches often struggle to maintain motion modeling accuracy an...
Learning Dynamic Scene Reconstruction with Sinusoidal Geometric Priors. We propose SirenPose, a novel loss function that combines the periodic activation properties of sinusoidal representation networks with geometric priors derived from keypoint structures to improve the accuracy of dynamic 3D scene reconstruction. Ex...
cv_vision
cs.CV
2025-12-25T20:51:19+00:00
2512.24016v1
FitControler: Toward Fit-Aware Virtual Try-On
Realistic virtual try-on (VTON) concerns not only faithful rendering of garment details but also coordination of the style. Prior art typically pursues the former, but neglects a key factor that shapes the holistic style -- garment fit. Garment fit delineates how a garment aligns with the body of a wearer and is a fund...
FitControler: Toward Fit-Aware Virtual Try-On. Realistic virtual try-on (VTON) concerns not only faithful rendering of garment details but also coordination of the style. Prior art typically pursues the former, but neglects a key factor that shapes the holistic style -- garment fit. Garment fit delineates how a garment...
cv_vision
cs.CV
2025-12-30T06:31:19+00:00
2510.24754v1
Certainty in Uncertainty: Reasoning over Uncertain Knowledge Graphs with Statistical Guarantees
Uncertain knowledge graph embedding (UnKGE) methods learn vector representations that capture both structural and uncertainty information to predict scores of unseen triples. However, existing methods produce only point estimates, without quantifying predictive uncertainty-limiting their reliability in high-stakes appl...
Certainty in Uncertainty: Reasoning over Uncertain Knowledge Graphs with Statistical Guarantees. Uncertain knowledge graph embedding (UnKGE) methods learn vector representations that capture both structural and uncertainty information to predict scores of unseen triples. However, existing methods produce only point est...
stat_ml
stat.ML
2025-10-18T17:58:17+00:00
2511.20122v1
Towards A Tri-View Diffusion Framework for Recommendation
Diffusion models (DMs) have recently gained significant interest for their exceptional potential in recommendation tasks. This stems primarily from their prominent capability in distilling, modeling, and generating comprehensive user preferences. However, previous work fails to examine DMs in recommendation tasks throu...
Towards A Tri-View Diffusion Framework for Recommendation. Diffusion models (DMs) have recently gained significant interest for their exceptional potential in recommendation tasks. This stems primarily from their prominent capability in distilling, modeling, and generating comprehensive user preferences. However, previ...
ir_retrieval
cs.IR
2025-11-25T09:43:00+00:00
2512.17308v1
Large Language Models as Pokémon Battle Agents: Strategic Play and Content Generation
Strategic decision-making in Pokémon battles presents a unique testbed for evaluating large language models. Pokémon battles demand reasoning about type matchups, statistical trade-offs, and risk assessment, skills that mirror human strategic thinking. This work examines whether Large Language Models (LLMs) can serve a...
Large Language Models as Pokémon Battle Agents: Strategic Play and Content Generation. Strategic decision-making in Pokémon battles presents a unique testbed for evaluating large language models. Pokémon battles demand reasoning about type matchups, statistical trade-offs, and risk assessment, skills that mirror human ...
ai_general
cs.AI
2025-12-19T07:46:29+00:00
2511.01448v2
LiCoMemory: Lightweight and Cognitive Agentic Memory for Efficient Long-Term Reasoning
Large Language Model (LLM) agents exhibit remarkable conversational and reasoning capabilities but remain constrained by limited context windows and the lack of persistent memory. Recent efforts address these limitations via external memory architectures, often employing graph-based representations, yet most adopt flat...
LiCoMemory: Lightweight and Cognitive Agentic Memory for Efficient Long-Term Reasoning. Large Language Model (LLM) agents exhibit remarkable conversational and reasoning capabilities but remain constrained by limited context windows and the lack of persistent memory. Recent efforts address these limitations via externa...
ir_retrieval
cs.IR
2025-11-03T11:02:40+00:00
2509.03910v1
An invertible generative model for forward and inverse problems
We formulate the inverse problem in a Bayesian framework and aim to train a generative model that allows us to simulate (i.e., sample from the likelihood) and do inference (i.e., sample from the posterior). We review the use of triangular normalizing flows for conditional sampling in this context and show how to combin...
An invertible generative model for forward and inverse problems. We formulate the inverse problem in a Bayesian framework and aim to train a generative model that allows us to simulate (i.e., sample from the likelihood) and do inference (i.e., sample from the posterior). We review the use of triangular normalizing flow...
stat_ml
stat.ML
2025-09-04T06:04:10+00:00
2512.17621v1
PathFLIP: Fine-grained Language-Image Pretraining for Versatile Computational Pathology
While Vision-Language Models (VLMs) have achieved notable progress in computational pathology (CPath), the gigapixel scale and spatial heterogeneity of Whole Slide Images (WSIs) continue to pose challenges for multimodal understanding. Existing alignment methods struggle to capture fine-grained correspondences between ...
PathFLIP: Fine-grained Language-Image Pretraining for Versatile Computational Pathology. While Vision-Language Models (VLMs) have achieved notable progress in computational pathology (CPath), the gigapixel scale and spatial heterogeneity of Whole Slide Images (WSIs) continue to pose challenges for multimodal understand...
cv_vision
cs.CV
2025-12-19T14:26:50+00:00
2511.07573v2
A Hybrid Multimodal Deep Learning Framework for Intelligent Fashion Recommendation
The rapid expansion of online fashion platforms has created an increasing demand for intelligent recommender systems capable of understanding both visual and textual cues. This paper proposes a hybrid multimodal deep learning framework for fashion recommendation that jointly addresses two key tasks: outfit compatibilit...
A Hybrid Multimodal Deep Learning Framework for Intelligent Fashion Recommendation. The rapid expansion of online fashion platforms has created an increasing demand for intelligent recommender systems capable of understanding both visual and textual cues. This paper proposes a hybrid multimodal deep learning framework ...
ir_retrieval
cs.IR
2025-11-10T19:33:12+00:00
2512.18412v1
Few-Shot Learning of a Graph-Based Neural Network Model Without Backpropagation
We propose a structural-graph approach to classifying contour images in a few-shot regime without using backpropagation. The core idea is to make structure the carrier of explanations: an image is encoded as an attributed graph (critical points and lines represented as nodes with geometric attributes), and generalizati...
Few-Shot Learning of a Graph-Based Neural Network Model Without Backpropagation. We propose a structural-graph approach to classifying contour images in a few-shot regime without using backpropagation. The core idea is to make structure the carrier of explanations: an image is encoded as an attributed graph (critical p...
ai_general
cs.AI
2025-12-20T16:23:51+00:00
2509.21160v1
WISER: Segmenting watermarked region - an epidemic change-point perspective
With the increasing popularity of large language models, concerns over content authenticity have led to the development of myriad watermarking schemes. These schemes can be used to detect a machine-generated text via an appropriate key, while being imperceptible to readers with no such keys. The corresponding detection...
WISER: Segmenting watermarked region - an epidemic change-point perspective. With the increasing popularity of large language models, concerns over content authenticity have led to the development of myriad watermarking schemes. These schemes can be used to detect a machine-generated text via an appropriate key, while ...
stat_ml
stat.ML
2025-09-25T13:44:34+00:00
2512.05464v1
Dynamic Alignment for Collective Agency: Toward a Scalable Self-Improving Framework for Open-Ended LLM Alignment
Large Language Models (LLMs) are typically aligned with human values using preference data or predefined principles such as helpfulness, honesty, and harmlessness. However, as AI systems progress toward Artificial General Intelligence (AGI) and Artificial Superintelligence (ASI), such value systems may become insuffici...
Dynamic Alignment for Collective Agency: Toward a Scalable Self-Improving Framework for Open-Ended LLM Alignment. Large Language Models (LLMs) are typically aligned with human values using preference data or predefined principles such as helpfulness, honesty, and harmlessness. However, as AI systems progress toward Art...
cl_nlp
cs.CL
2025-12-05T06:46:00+00:00
2512.19602v1
No Data? No Problem: Robust Vision-Tabular Learning with Missing Values
Large-scale medical biobanks provide imaging data complemented by extensive tabular information, such as demographics or clinical measurements. However, this abundance of tabular attributes does not reflect real-world datasets, where only a subset of attributes may be available. This discrepancy calls for methods that ...
No Data? No Problem: Robust Vision-Tabular Learning with Missing Values. Large-scale medical biobanks provide imaging data complemented by extensive tabular information, such as demographics or clinical measurements. However, this abundance of tabular attributes does not reflect real-world datasets, where only a subset...
cv_vision
cs.CV
2025-12-22T17:35:32+00:00
2512.15042v1
DASH: Dialogue-Aware Similarity and Handshake Recognition for Topic Segmentation in Public-Channel Conversations
Dialogue Topic Segmentation (DTS) is crucial for understanding task-oriented public-channel communications, such as maritime VHF dialogues, which feature informal speech and implicit transitions. To address the limitations of traditional methods, we propose DASH-DTS, a novel LLM-based framework. Its core contributions ...
DASH: Dialogue-Aware Similarity and Handshake Recognition for Topic Segmentation in Public-Channel Conversations. Dialogue Topic Segmentation (DTS) is crucial for understanding task-oriented public-channel communications, such as maritime VHF dialogues, which feature informal speech and implicit transitions. To address...
cl_nlp
cs.CL
2025-12-17T03:13:10+00:00
2512.23090v2
Benchmark Success, Clinical Failure: When Reinforcement Learning Optimizes for Benchmarks, Not Patients
Recent Reinforcement Learning (RL) advances for Large Language Models (LLMs) have improved reasoning tasks, yet their resource-constrained application to medical imaging remains underexplored. We introduce ChexReason, a vision-language model trained via R1-style methodology (SFT followed by GRPO) using only 2,000 SFT s...
Benchmark Success, Clinical Failure: When Reinforcement Learning Optimizes for Benchmarks, Not Patients. Recent Reinforcement Learning (RL) advances for Large Language Models (LLMs) have improved reasoning tasks, yet their resource-constrained application to medical imaging remains underexplored. We introduce ChexReaso...
ai_general
cs.AI
2025-12-28T21:57:42+00:00
2510.14704v1
Dataset Pruning in RecSys and ML: Best Practice or Mal-Practice?
Offline evaluations in recommender system research depend heavily on datasets, many of which are pruned, such as the widely used MovieLens collections. This thesis examines the impact of data pruning - specifically, removing users with fewer than a specified number of interactions - on both dataset characteristics and ...
Dataset Pruning in RecSys and ML: Best Practice or Mal-Practice?. Offline evaluations in recommender system research depend heavily on datasets, many of which are pruned, such as the widely used MovieLens collections. This thesis examines the impact of data pruning - specifically, removing users with fewer than a speci...
ir_retrieval
cs.IR
2025-10-16T14:08:30+00:00
2512.14980v3
Softly Constrained Denoisers for Diffusion Models
Diffusion models struggle to produce samples that respect constraints, a common requirement in scientific applications. Recent approaches have introduced regularization terms in the loss or guidance methods during sampling to enforce such constraints, but they bias the generative model away from the true data distribut...
Softly Constrained Denoisers for Diffusion Models. Diffusion models struggle to produce samples that respect constraints, a common requirement in scientific applications. Recent approaches have introduced regularization terms in the loss or guidance methods during sampling to enforce such constraints, but they bias the...
lg_ml
cs.LG
2025-12-17T00:35:45+00:00
2510.06838v1
Crossing Domains without Labels: Distant Supervision for Term Extraction
Automatic Term Extraction (ATE) is a critical component in downstream NLP tasks such as document tagging, ontology construction and patent analysis. Current state-of-the-art methods require expensive human annotation and struggle with domain transfer, limiting their practical deployment. This highlights the need for mo...
Crossing Domains without Labels: Distant Supervision for Term Extraction. Automatic Term Extraction (ATE) is a critical component in downstream NLP tasks such as document tagging, ontology construction and patent analysis. Current state-of-the-art methods require expensive human annotation and struggle with domain tran...
ir_retrieval
cs.IR
2025-10-08T10:02:40+00:00
2512.09114v2
AI TIPS 2.0: A Comprehensive Framework for Operationalizing AI Governance
The deployment of AI systems faces three critical governance challenges that current frameworks fail to adequately address. First, organizations struggle with inadequate risk assessment at the use case level, exemplified by the Humana class action lawsuit and other high impact cases where an AI system deployed to produ...
AI TIPS 2.0: A Comprehensive Framework for Operationalizing AI Governance. The deployment of AI systems faces three critical governance challenges that current frameworks fail to adequately address. First, organizations struggle with inadequate risk assessment at the use case level, exemplified by the Humana class acti...
ai_general
cs.AI
2025-12-09T20:57:22+00:00
2511.15146v1
Beyond Uncertainty Sets: Leveraging Optimal Transport to Extend Conformal Predictive Distribution to Multivariate Settings
Conformal prediction (CP) constructs uncertainty sets for model outputs with finite-sample coverage guarantees. A candidate output is included in the prediction set if its non-conformity score is not considered extreme relative to the scores observed on a set of calibration examples. However, this procedure is only str...
Beyond Uncertainty Sets: Leveraging Optimal Transport to Extend Conformal Predictive Distribution to Multivariate Settings. Conformal prediction (CP) constructs uncertainty sets for model outputs with finite-sample coverage guarantees. A candidate output is included in the prediction set if its non-conformity score is ...
stat_ml
stat.ML
2025-11-19T05:59:01+00:00
2512.23850v2
The Drill-Down and Fabricate Test (DDFT): A Protocol for Measuring Epistemic Robustness in Language Models
Current language model evaluations measure what models know under ideal conditions but not how robustly they know it under realistic stress. Static benchmarks like MMLU and TruthfulQA cannot distinguish a model that lacks knowledge from one whose verification mechanisms collapse when information degrades or adversaries...
The Drill-Down and Fabricate Test (DDFT): A Protocol for Measuring Epistemic Robustness in Language Models. Current language model evaluations measure what models know under ideal conditions but not how robustly they know it under realistic stress. Static benchmarks like MMLU and TruthfulQA cannot distinguish a model t...
ai_general
cs.AI
2025-12-29T20:29:09+00:00
2512.15614v1
Behavior Tokens Speak Louder: Disentangled Explainable Recommendation with Behavior Vocabulary
Recent advances in explainable recommendations have explored the integration of language models to analyze natural language rationales for user-item interactions. Despite their potential, existing methods often rely on ID-based representations that obscure semantic meaning and impose structural constraints on language ...
Behavior Tokens Speak Louder: Disentangled Explainable Recommendation with Behavior Vocabulary. Recent advances in explainable recommendations have explored the integration of language models to analyze natural language rationales for user-item interactions. Despite their potential, existing methods often rely on ID-ba...
lg_ml
cs.LG
2025-12-17T17:24:24+00:00
2512.12411v2
Detecting the Disturbance: A Nuanced View of Introspective Abilities in LLMs
Can large language models introspect, that is, accurately detect perturbations to their own internal states? We systematically investigate this question using activation steering in Meta-Llama-3.1-8B-Instruct. First, we show that the binary detection paradigm used in prior work conflates introspection with a methodolog...
Detecting the Disturbance: A Nuanced View of Introspective Abilities in LLMs. Can large language models introspect, that is, accurately detect perturbations to their own internal states? We systematically investigate this question using activation steering in Meta-Llama-3.1-8B-Instruct. First, we show that the binary d...
ai_general
cs.AI
2025-12-13T17:51:13+00:00
2512.17164v1
TCDE: Topic-Centric Dual Expansion of Queries and Documents with Large Language Models for Information Retrieval
Query Expansion (QE) enriches queries and Document Expansion (DE) enriches documents, and these two techniques are often applied separately. However, such separate application may lead to semantic misalignment between the expanded queries (or documents) and their relevant documents (or queries). To address this serious...
TCDE: Topic-Centric Dual Expansion of Queries and Documents with Large Language Models for Information Retrieval. Query Expansion (QE) enriches queries and Document Expansion (DE) enriches documents, and these two techniques are often applied separately. However, such separate application may lead to semantic misalignm...
ir_retrieval
cs.IR
2025-12-19T01:57:17+00:00
2512.05119v1
RAG-IGBench: Innovative Evaluation for RAG-based Interleaved Generation in Open-domain Question Answering
In real-world scenarios, providing user queries with visually enhanced responses can considerably benefit understanding and memory, underscoring the great value of interleaved image-text generation. Despite recent progress, like the visual autoregressive model that unifies text and image processing in a single transfor...
RAG-IGBench: Innovative Evaluation for RAG-based Interleaved Generation in Open-domain Question Answering. In real-world scenarios, providing user queries with visually enhanced responses can considerably benefit understanding and memory, underscoring the great value of interleaved image-text generation. Despite recent...
ir_retrieval
cs.IR
2025-10-11T03:06:39+00:00
2512.22615v2
Dream-VL & Dream-VLA: Open Vision-Language and Vision-Language-Action Models with Diffusion Language Model Backbone
While autoregressive Large Vision-Language Models (VLMs) have achieved remarkable success, their sequential generation often limits their efficacy in complex visual planning and dynamic robotic control. In this work, we investigate the potential of constructing Vision-Language Models upon diffusion-based large language...
Dream-VL & Dream-VLA: Open Vision-Language and Vision-Language-Action Models with Diffusion Language Model Backbone. While autoregressive Large Vision-Language Models (VLMs) have achieved remarkable success, their sequential generation often limits their efficacy in complex visual planning and dynamic robotic control. ...
cv_vision
cs.CV
2025-12-27T14:46:24+00:00
2512.20822v2
MediEval: A Unified Medical Benchmark for Patient-Contextual and Knowledge-Grounded Reasoning in LLMs
Large Language Models (LLMs) are increasingly applied to medicine, yet their adoption is limited by concerns over reliability and safety. Existing evaluations either test factual medical knowledge in isolation or assess patient-level reasoning without verifying correctness, leaving a critical gap. We introduce MediEval...
MediEval: A Unified Medical Benchmark for Patient-Contextual and Knowledge-Grounded Reasoning in LLMs. Large Language Models (LLMs) are increasingly applied to medicine, yet their adoption is limited by concerns over reliability and safety. Existing evaluations either test factual medical knowledge in isolation or asse...
cl_nlp
cs.CL
2025-12-23T22:52:24+00:00
2512.01231v1
Implicitly Normalized Online PCA: A Regularized Algorithm with Exact High-Dimensional Dynamics
Many online learning algorithms, including classical online PCA methods, enforce explicit normalization steps that discard the evolving norm of the parameter vector. We show that this norm can in fact encode meaningful information about the underlying statistical structure of the problem, and that exploiting this infor...
Implicitly Normalized Online PCA: A Regularized Algorithm with Exact High-Dimensional Dynamics. Many online learning algorithms, including classical online PCA methods, enforce explicit normalization steps that discard the evolving norm of the parameter vector. We show that this norm can in fact encode meaningful infor...
stat_ml
stat.ML
2025-12-01T03:17:29+00:00
2512.10430v1
T-pro 2.0: An Efficient Russian Hybrid-Reasoning Model and Playground
We introduce T-pro 2.0, an open-weight Russian LLM for hybrid reasoning and efficient inference. The model supports direct answering and reasoning-trace generation, using a Cyrillic-dense tokenizer and an adapted EAGLE speculative-decoding pipeline to reduce latency. To enable reproducible and extensible research, we r...
T-pro 2.0: An Efficient Russian Hybrid-Reasoning Model and Playground. We introduce T-pro 2.0, an open-weight Russian LLM for hybrid reasoning and efficient inference. The model supports direct answering and reasoning-trace generation, using a Cyrillic-dense tokenizer and an adapted EAGLE speculative-decoding pipeline ...
cl_nlp
cs.CL
2025-12-11T08:40:10+00:00
2510.25240v3
Generative Bayesian Optimization: Generative Models as Acquisition Functions
We present a general strategy for turning generative models into candidate solution samplers for batch Bayesian optimization (BO). The use of generative models for BO enables large batch scaling as generative sampling, optimization of non-continuous design spaces, and high-dimensional and combinatorial design. Inspired...
Generative Bayesian Optimization: Generative Models as Acquisition Functions. We present a general strategy for turning generative models into candidate solution samplers for batch Bayesian optimization (BO). The use of generative models for BO enables large batch scaling as generative sampling, optimization of non-con...
stat_ml
stat.ML
2025-10-29T07:42:25+00:00
2512.22279v1
Hierarchical Stacking Optimization Using Dirichlet's Process (SoDip): Towards Accelerated Design for Graft Polymerization
Radiation-induced grafting (RIG) enables precise functionalization of polymer films for ion-exchange membranes, CO2-separation membranes, and battery electrolytes by generating radicals on robust substrates to graft desired monomers. However, reproducibility remains limited due to unreported variability in base-film mo...
Hierarchical Stacking Optimization Using Dirichlet's Process (SoDip): Towards Accelerated Design for Graft Polymerization. Radiation-induced grafting (RIG) enables precise functionalization of polymer films for ion-exchange membranes, CO2-separation membranes, and battery electrolytes by generating radicals on robust s...
lg_ml
cs.LG
2025-12-25T05:36:26+00:00
2512.07796v1
Large Causal Models from Large Language Models
We introduce a new paradigm for building large causal models (LCMs) that exploits the enormous potential latent in today's large language models (LLMs). We describe our ongoing experiments with an implemented system called DEMOCRITUS (Decentralized Extraction of Manifold Ontologies of Causal Relations Integrating Topos...
Large Causal Models from Large Language Models. We introduce a new paradigm for building large causal models (LCMs) that exploits the enormous potential latent in today's large language models (LLMs). We describe our ongoing experiments with an implemented system called DEMOCRITUS (Decentralized Extraction of Manifold ...
ai_general
cs.AI
2025-12-08T18:28:04+00:00
2511.11255v2
Align$^3$GR: Unified Multi-Level Alignment for LLM-based Generative Recommendation
Large Language Models (LLMs) demonstrate significant advantages in leveraging structured world knowledge and multi-step reasoning capabilities. However, fundamental challenges arise when transforming LLMs into real-world recommender systems due to semantic and behavioral misalignment. To bridge this gap, we propose Ali...
Align$^3$GR: Unified Multi-Level Alignment for LLM-based Generative Recommendation. Large Language Models (LLMs) demonstrate significant advantages in leveraging structured world knowledge and multi-step reasoning capabilities. However, fundamental challenges arise when transforming LLMs into real-world recommender sys...
ir_retrieval
cs.IR
2025-11-14T12:52:43+00:00
2512.16334v5
Pretrained battery transformer (PBT): A foundation model for universal battery life prediction
Early prediction of battery cycle life is essential for improving battery design, manufacturing, and deployment. However, despite encouraging results with machine learning, progress remains constrained by scarce data and data heterogeneity across battery chemistries, specifications, formation protocols, and operating c...
Pretrained battery transformer (PBT): A foundation model for universal battery life prediction. Early prediction of battery cycle life is essential for improving battery design, manufacturing, and deployment. However, despite encouraging results with machine learning, progress remains constrained by scarce data and dat...
lg_ml
cs.LG
2025-12-18T09:17:45+00:00
2512.17986v1
FedOAED: Federated On-Device Autoencoder Denoiser for Heterogeneous Data under Limited Client Availability
Over the last few decades, machine learning (ML) and deep learning (DL) solutions have demonstrated their potential across many applications by leveraging large amounts of high-quality data. However, strict data-sharing regulations such as the General Data Protection Regulation (GDPR) and the Health Insurance Portabili...
FedOAED: Federated On-Device Autoencoder Denoiser for Heterogeneous Data under Limited Client Availability. Over the last few decades, machine learning (ML) and deep learning (DL) solutions have demonstrated their potential across many applications by leveraging large amounts of high-quality data. However, strict data-...
lg_ml
cs.LG
2025-12-19T15:35:48+00:00
2512.19320v1
MAGIC: Achieving Superior Model Merging via Magnitude Calibration
The proliferation of pre-trained models has given rise to a wide array of specialised, fine-tuned models. Model merging aims to merge the distinct capabilities of these specialised models into a unified model, requiring minimal or even no additional training. A core objective of model merging is to ensure the merged mo...
MAGIC: Achieving Superior Model Merging via Magnitude Calibration. The proliferation of pre-trained models has given rise to a wide array of specialised, fine-tuned models. Model merging aims to merge the distinct capabilities of these specialised models into a unified model, requiring minimal or even no additional tra...
lg_ml
cs.LG
2025-12-22T12:13:17+00:00
2510.01698v3
TalkPlay-Tools: Conversational Music Recommendation with LLM Tool Calling
While the recent developments in large language models (LLMs) have successfully enabled generative recommenders with natural language interactions, their recommendation behavior is limited, leaving other simpler yet crucial components such as metadata or attribute filtering underutilized in the system. We propose an LL...
TalkPlay-Tools: Conversational Music Recommendation with LLM Tool Calling. While the recent developments in large language models (LLMs) have successfully enabled generative recommenders with natural language interactions, their recommendation behavior is limited, leaving other simpler yet crucial components such as me...
ir_retrieval
cs.IR
2025-10-02T06:08:54+00:00
2512.20606v1
Repurposing Video Diffusion Transformers for Robust Point Tracking
Point tracking aims to localize corresponding points across video frames, serving as a fundamental task for 4D reconstruction, robotics, and video editing. Existing methods commonly rely on shallow convolutional backbones such as ResNet that process frames independently, lacking temporal coherence and producing unrelia...
Repurposing Video Diffusion Transformers for Robust Point Tracking. Point tracking aims to localize corresponding points across video frames, serving as a fundamental task for 4D reconstruction, robotics, and video editing. Existing methods commonly rely on shallow convolutional backbones such as ResNet that process fr...
cv_vision
cs.CV
2025-12-23T18:54:10+00:00
2512.17121v1
The Effect of Negation on CLIP in Medical Imaging: Limitations of Contrastive Language-Image Pretraining
Large vision-language models like CLIP are increasingly used in medical imaging tasks due to their ability to align images and text without the need for extensive labeled data. This makes them particularly useful for applications like image retrieval, report generation, and classification in clinical settings. A potent...
The Effect of Negation on CLIP in Medical Imaging: Limitations of Contrastive Language-Image Pretraining. Large vision-language models like CLIP are increasingly used in medical imaging tasks due to their ability to align images and text without the need for extensive labeled data. This makes them particularly useful f...
lg_ml
cs.LG
2025-12-18T23:19:19+00:00
2512.19954v1
HistoWAS: A Pathomics Framework for Large-Scale Feature-Wide Association Studies of Tissue Topology and Patient Outcomes
High-throughput "pathomic" analysis of Whole Slide Images (WSIs) offers new opportunities to study tissue characteristics and for biomarker discovery. However, the clinical relevance of the tissue characteristics at the micro- and macro-environment level is limited by the lack of tools that facilitate the measurement o...
HistoWAS: A Pathomics Framework for Large-Scale Feature-Wide Association Studies of Tissue Topology and Patient Outcomes. High-throughput "pathomic" analysis of Whole Slide Images (WSIs) offers new opportunities to study tissue characteristics and for biomarker discovery. However, the clinical relevance of the tissue c...
cv_vision
cs.CV
2025-12-23T00:58:27+00:00
2601.00887v1
VideoCuRL: Video Curriculum Reinforcement Learning with Orthogonal Difficulty Decomposition
Reinforcement Learning (RL) is crucial for empowering VideoLLMs with complex spatiotemporal reasoning. However, current RL paradigms predominantly rely on random data shuffling or naive curriculum strategies based on scalar difficulty metrics. We argue that scalar metrics fail to disentangle two orthogonal challenges i...
VideoCuRL: Video Curriculum Reinforcement Learning with Orthogonal Difficulty Decomposition. Reinforcement Learning (RL) is crucial for empowering VideoLLMs with complex spatiotemporal reasoning. However, current RL paradigms predominantly rely on random data shuffling or naive curriculum strategies based on scalar dif...
cv_vision
cs.CV
2025-12-31T09:25:36+00:00
2512.23594v1
Detection Fire in Camera RGB-NIR
Improving the accuracy of fire detection using infrared night vision cameras remains a challenging task. Previous studies have reported strong performance with popular detection models. For example, YOLOv7 achieved an mAP50-95 of 0.51 using an input image size of 640 x 1280, RT-DETR reached an mAP50-95 of 0.65 with an ...
Detection Fire in Camera RGB-NIR. Improving the accuracy of fire detection using infrared night vision cameras remains a challenging task. Previous studies have reported strong performance with popular detection models. For example, YOLOv7 achieved an mAP50-95 of 0.51 using an input image size of 640 x 1280, RT-DETR re...
cv_vision
cs.CV
2025-12-29T16:48:24+00:00
2511.03963v1
Robust inference using density-powered Stein operators
We introduce a density-power weighted variant for the Stein operator, called the $γ$-Stein operator. This is a novel class of operators derived from the $γ$-divergence, designed to build robust inference methods for unnormalized probability models. The operator's construction (weighting by the model density raised to a...
Robust inference using density-powered Stein operators. We introduce a density-power weighted variant for the Stein operator, called the $γ$-Stein operator. This is a novel class of operators derived from the $γ$-divergence, designed to build robust inference methods for unnormalized probability models. The operator's ...
stat_ml
stat.ML
2025-11-06T01:32:17+00:00
2511.08808v1
Effects of label noise on the classification of outlier observations
This study investigates the impact of adding noise to the training set classes in classification tasks using the BCOPS algorithm (Balanced and Conformal Optimized Prediction Sets), proposed by Guan & Tibshirani (2022). The BCOPS algorithm is an application of conformal prediction combined with a machine learning method...
Effects of label noise on the classification of outlier observations. This study investigates the impact of adding noise to the training set classes in classification tasks using the BCOPS algorithm (Balanced and Conformal Optimized Prediction Sets), proposed by Guan & Tibshirani (2022). The BCOPS algorithm is an appli...
stat_ml
stat.ML
2025-11-11T22:22:54+00:00
2512.24002v1
Tracing the Heart's Pathways: ECG Representation Learning from a Cardiac Conduction Perspective
The multi-lead electrocardiogram (ECG) stands as a cornerstone of cardiac diagnosis. Recent strides in electrocardiogram self-supervised learning (eSSL) have brightened prospects for enhancing representation learning without relying on high-quality annotations. Yet earlier eSSL methods suffer a key limitation: they foc...
Tracing the Heart's Pathways: ECG Representation Learning from a Cardiac Conduction Perspective. The multi-lead electrocardiogram (ECG) stands as a cornerstone of cardiac diagnosis. Recent strides in electrocardiogram self-supervised learning (eSSL) have brightened prospects for enhancing representation learning withou...
lg_ml
cs.LG
2025-12-30T05:46:48+00:00
2512.10903v1
Multi-Granular Node Pruning for Circuit Discovery
Circuit discovery aims to identify minimal subnetworks that are responsible for specific behaviors in large language models (LLMs). Existing approaches primarily rely on iterative edge pruning, which is computationally expensive and limited to coarse-grained units such as attention heads or MLP blocks, overlooking fine...
Multi-Granular Node Pruning for Circuit Discovery. Circuit discovery aims to identify minimal subnetworks that are responsible for specific behaviors in large language models (LLMs). Existing approaches primarily rely on iterative edge pruning, which is computationally expensive and limited to coarse-grained units such...
ai_general
cs.AI
2025-12-11T18:32:15+00:00
2601.00121v1
Ask, Clarify, Optimize: Human-LLM Agent Collaboration for Smarter Inventory Control
Inventory management remains a challenge for many small and medium-sized businesses that lack the expertise to deploy advanced optimization methods. This paper investigates whether Large Language Models (LLMs) can help bridge this gap. We show that employing LLMs as direct, end-to-end solvers incurs a significant "hall...
Ask, Clarify, Optimize: Human-LLM Agent Collaboration for Smarter Inventory Control. Inventory management remains a challenge for many small and medium-sized businesses that lack the expertise to deploy advanced optimization methods. This paper investigates whether Large Language Models (LLMs) can help bridge this gap....
ai_general
cs.AI
2025-12-31T21:45:54+00:00
2509.19276v1
A Gradient Flow Approach to Solving Inverse Problems with Latent Diffusion Models
Solving ill-posed inverse problems requires powerful and flexible priors. We propose leveraging pretrained latent diffusion models for this task through a new training-free approach, termed Diffusion-regularized Wasserstein Gradient Flow (DWGF). Specifically, we formulate the posterior sampling problem as a regularized...
A Gradient Flow Approach to Solving Inverse Problems with Latent Diffusion Models. Solving ill-posed inverse problems requires powerful and flexible priors. We propose leveraging pretrained latent diffusion models for this task through a new training-free approach, termed Diffusion-regularized Wasserstein Gradient Flow...
stat_ml
stat.ML
2025-09-23T17:41:43+00:00
2512.09675v3
d-TreeRPO: Towards More Reliable Policy Optimization for Diffusion Language Models
Reinforcement learning (RL) is pivotal for enhancing the reasoning capabilities of diffusion large language models (dLLMs). However, existing dLLM policy optimization methods suffer from two critical reliability bottlenecks: (1) reward sparsity, arising from coarse or unverifiable signals that impede accurate advantage...
d-TreeRPO: Towards More Reliable Policy Optimization for Diffusion Language Models. Reinforcement learning (RL) is pivotal for enhancing the reasoning capabilities of diffusion large language models (dLLMs). However, existing dLLM policy optimization methods suffer from two critical reliability bottlenecks: (1) reward ...
cl_nlp
cs.CL
2025-12-10T14:20:07+00:00
2512.15959v1
BRAID: Bounded Reasoning for Autonomous Inference and Decisions
Large Language Models (LLMs) exhibit nonlinear relationships between performance, cost, and token usage. This paper presents a quantitative study on structured prompting using BRAID (Bounded Reasoning for Au tonomous Inference and Decisions) across multiple GPT model tiers, eval uated on the AdvancedIF, GSM-Hard, and t...
BRAID: Bounded Reasoning for Autonomous Inference and Decisions. Large Language Models (LLMs) exhibit nonlinear relationships between performance, cost, and token usage. This paper presents a quantitative study on structured prompting using BRAID (Bounded Reasoning for Au tonomous Inference and Decisions) across multip...
cl_nlp
cs.CL
2025-12-17T20:46:44+00:00
2511.18347v1
Time Matters: Enhancing Sequential Recommendations with Time-Guided Graph Neural ODEs
Sequential recommendation (SR) is widely deployed in e-commerce platforms, streaming services, etc., revealing significant potential to enhance user experience. However, existing methods often overlook two critical factors: irregular user interests between interactions and highly uneven item distributions over time. Th...
Time Matters: Enhancing Sequential Recommendations with Time-Guided Graph Neural ODEs. Sequential recommendation (SR) is widely deployed in e-commerce platforms, streaming services, etc., revealing significant potential to enhance user experience. However, existing methods often overlook two critical factors: irregular...
ir_retrieval
cs.IR
2025-11-23T08:46:27+00:00
2512.22031v1
From In Silico to In Vitro: Evaluating Molecule Generative Models for Hit Generation
Hit identification is a critical yet resource-intensive step in the drug discovery pipeline, traditionally relying on high-throughput screening of large compound libraries. Despite advancements in virtual screening, these methods remain time-consuming and costly. Recent progress in deep learning has enabled the develop...
From In Silico to In Vitro: Evaluating Molecule Generative Models for Hit Generation. Hit identification is a critical yet resource-intensive step in the drug discovery pipeline, traditionally relying on high-throughput screening of large compound libraries. Despite advancements in virtual screening, these methods rema...
lg_ml
cs.LG
2025-12-26T14:02:59+00:00
2512.22274v2
GeCo: Evaluating Geometric Consistency for Video Generation via Motion and Structure
We introduce GeCo, a geometry-grounded metric for jointly detecting geometric deformation and occlusion-inconsistency artifacts in static scenes. By fusing residual motion and depth priors, GeCo produces interpretable, dense consistency maps that reveal these artifacts. We use GeCo to systematically benchmark recent vi...
GeCo: Evaluating Geometric Consistency for Video Generation via Motion and Structure. We introduce GeCo, a geometry-grounded metric for jointly detecting geometric deformation and occlusion-inconsistency artifacts in static scenes. By fusing residual motion and depth priors, GeCo produces interpretable, dense consisten...
cv_vision
cs.CV
2025-12-25T03:28:28+00:00
2512.22877v1
M-ErasureBench: A Comprehensive Multimodal Evaluation Benchmark for Concept Erasure in Diffusion Models
Text-to-image diffusion models may generate harmful or copyrighted content, motivating research on concept erasure. However, existing approaches primarily focus on erasing concepts from text prompts, overlooking other input modalities that are increasingly critical in real-world applications such as image editing and p...
M-ErasureBench: A Comprehensive Multimodal Evaluation Benchmark for Concept Erasure in Diffusion Models. Text-to-image diffusion models may generate harmful or copyrighted content, motivating research on concept erasure. However, existing approaches primarily focus on erasing concepts from text prompts, overlooking oth...
cv_vision
cs.CV
2025-12-28T10:58:36+00:00
2512.07631v1
The Agent Capability Problem: Predicting Solvability Through Information-Theoretic Bounds
When should an autonomous agent commit resources to a task? We introduce the Agent Capability Problem (ACP), a framework for predicting whether an agent can solve a problem under resource constraints. Rather than relying on empirical heuristics, ACP frames problem-solving as information acquisition: an agent requires $...
The Agent Capability Problem: Predicting Solvability Through Information-Theoretic Bounds. When should an autonomous agent commit resources to a task? We introduce the Agent Capability Problem (ACP), a framework for predicting whether an agent can solve a problem under resource constraints. Rather than relying on empir...
ai_general
cs.AI
2025-12-08T15:21:52+00:00
2512.05530v1
MIND: Multi-rationale INtegrated Discriminative Reasoning Framework for Multi-modal Large Models
Recently, multimodal large language models (MLLMs) have been widely applied to reasoning tasks. However, they suffer from limited multi-rationale semantic modeling, insufficient logical robustness, and are susceptible to misleading interpretations in complex scenarios. Therefore, we propose a Multi-rationale INtegrated...
MIND: Multi-rationale INtegrated Discriminative Reasoning Framework for Multi-modal Large Models. Recently, multimodal large language models (MLLMs) have been widely applied to reasoning tasks. However, they suffer from limited multi-rationale semantic modeling, insufficient logical robustness, and are susceptible to m...
ai_general
cs.AI
2025-12-05T08:41:44+00:00
2512.05863v1
Optimizing Medical Question-Answering Systems: A Comparative Study of Fine-Tuned and Zero-Shot Large Language Models with RAG Framework
Medical question-answering (QA) systems can benefit from advances in large language models (LLMs), but directly applying LLMs to the clinical domain poses challenges such as maintaining factual accuracy and avoiding hallucinations. In this paper, we present a retrieval-augmented generation (RAG) based medical QA system...
Optimizing Medical Question-Answering Systems: A Comparative Study of Fine-Tuned and Zero-Shot Large Language Models with RAG Framework. Medical question-answering (QA) systems can benefit from advances in large language models (LLMs), but directly applying LLMs to the clinical domain poses challenges such as maintaini...
cl_nlp
cs.CL
2025-12-05T16:38:47+00:00
2512.20937v1
Beyond Artifacts: Real-Centric Envelope Modeling for Reliable AI-Generated Image Detection
The rapid progress of generative models has intensified the need for reliable and robust detection under real-world conditions. However, existing detectors often overfit to generator-specific artifacts and remain highly sensitive to real-world degradations. As generative architectures evolve and images undergo multi-ro...
Beyond Artifacts: Real-Centric Envelope Modeling for Reliable AI-Generated Image Detection. The rapid progress of generative models has intensified the need for reliable and robust detection under real-world conditions. However, existing detectors often overfit to generator-specific artifacts and remain highly sensitiv...
cv_vision
cs.CV
2025-12-24T04:41:04+00:00
2512.18399v1
AraToken: Optimizing Arabic Tokenization with Normalization Pipeline and Language Extension for Qwen3
Tokenization is a critical preprocessing step for large language models (LLMs), directly impacting training efficiency and downstream performance. General-purpose tokenizers trained predominantly on English and Latin-script languages exhibit suboptimal performance on morphologically rich languages such as Arabic, resul...
AraToken: Optimizing Arabic Tokenization with Normalization Pipeline and Language Extension for Qwen3. Tokenization is a critical preprocessing step for large language models (LLMs), directly impacting training efficiency and downstream performance. General-purpose tokenizers trained predominantly on English and Latin-...
cl_nlp
cs.CL
2025-12-20T15:32:10+00:00
2512.03571v1
EnCompass: Enhancing Agent Programming with Search Over Program Execution Paths
We introduce a new approach to agent programming, the development of LLM-based agents. Current approaches to agent programming often entangle two aspects of agent design: the core workflow logic and the inference-time strategy (e.g., tree search). We introduce "probabilistic angelic nondeterminism" ("PAN"), a programmi...
EnCompass: Enhancing Agent Programming with Search Over Program Execution Paths. We introduce a new approach to agent programming, the development of LLM-based agents. Current approaches to agent programming often entangle two aspects of agent design: the core workflow logic and the inference-time strategy (e.g., tree ...
ai_general
cs.AI
2025-12-03T08:50:16+00:00
2510.08095v2
Beyond Real Data: Synthetic Data through the Lens of Regularization
Synthetic data can improve generalization when real data is scarce, but excessive reliance may introduce distributional mismatches that degrade performance. In this paper, we present a learning-theoretic framework to quantify the trade-off between synthetic and real data. Our approach leverages algorithmic stability to...
Beyond Real Data: Synthetic Data through the Lens of Regularization. Synthetic data can improve generalization when real data is scarce, but excessive reliance may introduce distributional mismatches that degrade performance. In this paper, we present a learning-theoretic framework to quantify the trade-off between syn...
stat_ml
stat.ML
2025-10-09T11:33:09+00:00
2512.12177v1
Floorplan2Guide: LLM-Guided Floorplan Parsing for BLV Indoor Navigation
Indoor navigation remains a critical challenge for people with visual impairments. The current solutions mainly rely on infrastructure-based systems, which limit their ability to navigate safely in dynamic environments. We propose a novel navigation approach that utilizes a foundation model to transform floor plans int...
Floorplan2Guide: LLM-Guided Floorplan Parsing for BLV Indoor Navigation. Indoor navigation remains a critical challenge for people with visual impairments. The current solutions mainly rely on infrastructure-based systems, which limit their ability to navigate safely in dynamic environments. We propose a novel navigati...
ai_general
cs.AI
2025-12-13T04:49:26+00:00
2512.21863v2
Frozen LVLMs for Micro-Video Recommendation: A Systematic Study of Feature Extraction and Fusion
Frozen Large Video Language Models (LVLMs) are increasingly employed in micro-video recommendation due to their strong multimodal understanding. However, their integration lacks systematic empirical evaluation: practitioners typically deploy LVLMs as fixed black-box feature extractors without systematically comparing a...
Frozen LVLMs for Micro-Video Recommendation: A Systematic Study of Feature Extraction and Fusion. Frozen Large Video Language Models (LVLMs) are increasingly employed in micro-video recommendation due to their strong multimodal understanding. However, their integration lacks systematic empirical evaluation: practitione...
ir_retrieval
cs.IR
2025-12-26T04:56:28+00:00
2512.21003v2
MVInverse: Feed-forward Multi-view Inverse Rendering in Seconds
Multi-view inverse rendering aims to recover geometry, materials, and illumination consistently across multiple viewpoints. When applied to multi-view images, existing single-view approaches often ignore cross-view relationships, leading to inconsistent results. In contrast, multi-view optimization methods rely on slow...
MVInverse: Feed-forward Multi-view Inverse Rendering in Seconds. Multi-view inverse rendering aims to recover geometry, materials, and illumination consistently across multiple viewpoints. When applied to multi-view images, existing single-view approaches often ignore cross-view relationships, leading to inconsistent r...
cv_vision
cs.CV
2025-12-24T06:59:29+00:00
2512.10100v2
Robust AI Security and Alignment: A Sisyphean Endeavor?
This manuscript establishes information-theoretic limitations for robustness of AI security and alignment by extending Gödel's incompleteness theorem to AI. Knowing these limitations and preparing for the challenges they bring is critically important for the responsible adoption of the AI technology. Practical approach...
Robust AI Security and Alignment: A Sisyphean Endeavor?. This manuscript establishes information-theoretic limitations for robustness of AI security and alignment by extending Gödel's incompleteness theorem to AI. Knowing these limitations and preparing for the challenges they bring is critically important for the resp...
ai_general
cs.AI
2025-12-10T21:44:10+00:00
2512.20350v1
Field-Space Attention for Structure-Preserving Earth System Transformers
Accurate and physically consistent modeling of Earth system dynamics requires machine-learning architectures that operate directly on continuous geophysical fields and preserve their underlying geometric structure. Here we introduce Field-Space attention, a mechanism for Earth system Transformers that computes attentio...
Field-Space Attention for Structure-Preserving Earth System Transformers. Accurate and physically consistent modeling of Earth system dynamics requires machine-learning architectures that operate directly on continuous geophysical fields and preserve their underlying geometric structure. Here we introduce Field-Space a...
lg_ml
cs.LG
2025-12-23T13:31:21+00:00
2512.24933v2
ADOPT: Adaptive Dependency-Guided Joint Prompt Optimization for Multi-Step LLM Pipelines
Multi-step LLM pipelines can solve complex tasks, but jointly optimizing prompts across steps remains challenging due to missing step-level supervision and inter-step dependency. We propose ADOPT, an adaptive dependency-guided joint prompt optimization framework for multi-step LLM pipelines. ADOPT analyzes the dependen...
ADOPT: Adaptive Dependency-Guided Joint Prompt Optimization for Multi-Step LLM Pipelines. Multi-step LLM pipelines can solve complex tasks, but jointly optimizing prompts across steps remains challenging due to missing step-level supervision and inter-step dependency. We propose ADOPT, an adaptive dependency-guided joi...
cl_nlp
cs.CL
2025-12-31T15:46:37+00:00
2512.05449v1
The Seeds of Scheming: Weakness of Will in the Building Blocks of Agentic Systems
Large language models display a peculiar form of inconsistency: they "know" the correct answer but fail to act on it. In human philosophy, this tension between global judgment and local impulse is called akrasia, or weakness of will. We propose akrasia as a foundational concept for analyzing inconsistency and goal drif...
The Seeds of Scheming: Weakness of Will in the Building Blocks of Agentic Systems. Large language models display a peculiar form of inconsistency: they "know" the correct answer but fail to act on it. In human philosophy, this tension between global judgment and local impulse is called akrasia, or weakness of will. We ...
ai_general
cs.AI
2025-12-05T05:57:40+00:00
2512.21583v1
A Medical Multimodal Diagnostic Framework Integrating Vision-Language Models and Logic Tree Reasoning
With the rapid growth of large language models (LLMs) and vision-language models (VLMs) in medicine, simply integrating clinical text and medical imaging does not guarantee reliable reasoning. Existing multimodal models often produce hallucinations or inconsistent chains of thought, limiting clinical trust. We propose ...
A Medical Multimodal Diagnostic Framework Integrating Vision-Language Models and Logic Tree Reasoning. With the rapid growth of large language models (LLMs) and vision-language models (VLMs) in medicine, simply integrating clinical text and medical imaging does not guarantee reliable reasoning. Existing multimodal mode...
ai_general
cs.AI
2025-12-25T09:01:06+00:00