AI研究論文每日精選

Mutarjim: Advancing Bidirectional Arabic-English Translation with a Small Language Model

Shifting AI Efficiency From Model-Centric to Data-Centric Compression

Alchemist: Turning Public Text-to-Image Data into Generative Gold

BizFinBench: A Business-Driven Real-World Financial Benchmark for Evaluating LLMs

PATS: Process-Level Adaptive Thinking Mode Switching

Embodied Agents Meet Personalization: Exploring Memory Utilization for Personalized Assistance

ARM: Adaptive Reasoning Model

Enigmata: Scaling Logical Reasoning in Large Language Models with Synthetic Verifiable Puzzles

Deciphering Trajectory-Aided LLM Reasoning: An Optimization Perspective

Surrogate Signals from Format and Length: Reinforcement Learning for Solving Mathematical Problems without Ground Truth Answers

B-score: Detecting biases in large language models using response history

Flex-Judge: Think Once, Judge Anywhere

MOOSE-Chem2: Exploring LLM Limits in Fine-Grained Scientific Hypothesis Discovery via Hierarchical Search

Learning to Reason without External Rewards

Can MLLMs Guide Me Home? A Benchmark Study on Fine-Grained Visual Reasoning from Transit Maps

Lifelong Safety Alignment for Language Models

Jodi: Unification of Visual Generation and Understanding via Joint Modeling

StructEval: Benchmarking LLMs' Capabilities to Generate Structural Outputs

Reinforcement Fine-Tuning Powers Reasoning Capability of Multimodal Large Language Models

REARANK: Reasoning Re-ranking Agent via Reinforcement Learning

ModernGBERT: German-only 1B Encoder Model Trained from Scratch

Discrete Markov Bridge

Omni-R1: Reinforcement Learning for Omnimodal Reasoning via Two-System Collaboration

Which Data Attributes Stimulate Math and Code Reasoning? An Investigation via Influence Functions

Hybrid Neural-MPM for Interactive Fluid Simulations in Real-Time

Vibe Coding vs. Agentic Coding: Fundamentals and Practical Implications of Agentic AI

Done Is Better than Perfect: Unlocking Efficient Reasoning by Structured Multi-Turn Decomposition

Memory-Efficient Visual Autoregressive Modeling with Scale-Aware KV Cache Compression

AdaCtrl: Towards Adaptive and Controllable Reasoning via Difficulty-Aware Budgeting

The Quest for Efficient Reasoning: A Data-Centric Benchmark to CoT Distillation

Hard Negative Contrastive Learning for Fine-Grained Geometric Understanding in Large Multimodal Models

Force Prompting: Video Generation Models Can Learn and Generalize Physics-based Control Signals

G1: Bootstrapping Perception and Reasoning Abilities of Vision-Language Model via Reinforcement Learning

Interleaved Reasoning for Large Language Models via Reinforcement Learning

WHISTRESS: Enriching Transcriptions with Sentence Stress Detection

WINA: Weight Informed Neuron Activation for Accelerating Large Language Model Inference

From Tens of Hours to Tens of Thousands: Scaling Back-Translation for Speech Recognition

MLR-Bench: Evaluating AI Agents on Open-Ended Machine Learning Research

LLaDA 1.5: Variance-Reduced Preference Optimization for Large Language Diffusion Models

STAR-R1: Spatial TrAnsformation Reasoning by Reinforcing Multimodal LLMs

InfantAgent-Next: A Multimodal Generalist Agent for Automated Computer Interaction

The Coverage Principle: A Framework for Understanding Compositional Generalization

Strong Membership Inference Attacks on Massive Datasets and (Moderately) Large Language Models

Dynamic Risk Assessments for Offensive Cybersecurity Agents

Accelerating Nash Learning from Human Feedback via Mirror Prox

Rethinking the Sampling Criteria in Reinforcement Learning for LLM Reasoning: A Competence-Difficulty Alignment Perspective

Position: Mechanistic Interpretability Should Prioritize Feature Consistency in SAEs

Don't "Overthink" Passage Reranking: Is Reasoning Truly Necessary?

An Embarrassingly Simple Defense Against LLM Abliteration Attacks

Hybrid Latent Reasoning via Reinforcement Learning

Bridging Supervised Learning and Reinforcement Learning in Math Reasoning

GLEAM: Learning Generalizable Exploration Policy for Active Mapping in Complex 3D Indoor Scenes

Error Typing for Smarter Rewards: Improving Process Reward Models with Error-Aware Hierarchical Supervision

DoctorAgent-RL: A Multi-Agent Collaborative Reinforcement Learning System for Multi-Turn Clinical Dialogue

Architectural Backdoors for Within-Batch Data Stealing and Model Inference Manipulation

UFT: Unifying Supervised and Reinforcement Fine-Tuning

EquivPruner: Boosting Efficiency and Quality in LLM-Based Search via Action Pruning

DiSA: Diffusion Step Annealing in Autoregressive Image Generation

Seeing is Believing, but How Much? A Comprehensive Analysis of Verbalized Calibration in Vision-Language Models

FLAME-MoE: A Transparent End-to-End Research Platform for Mixture-of-Experts Language Models

MMIG-Bench: Towards Comprehensive and Explainable Evaluation of Multi-Modal Image Generation Models

The Pragmatic Mind of Machines: Tracing the Emergence of Pragmatic Competence in Large Language Models

InstructPart: Task-Oriented Part Segmentation with Instruction Reasoning

TAGS: A Test-Time Generalist-Specialist Framework with Retrieval-Augmented Reasoning and Verification

Towards Holistic Evaluation of Large Audio-Language Models: A Comprehensive Survey

EgoZero: Robot Learning from Smart Glasses

MOLE: Metadata Extraction and Validation in Scientific Papers Using LLMs

The Birth of Knowledge: Emergent Features across Time, Space, and Scale in Large Language Models

CASS: Nvidia to AMD Transpilation with Data, Models, and Benchmark

Textual Steering Vectors Can Improve Visual Understanding in Multimodal Large Language Models

Option-aware Temporally Abstracted Value for Offline Goal-Conditioned Reinforcement Learning