AI 论文速览
最近采集到 163 篇论文，最新在前。
- cs.AI DreamX: Video Diffusion Model is a Universal Latent Action Learner
- cs.AI From Autonomous Agents to Cognitive Symbionts: Toward AI-Augmented Human Intelligence
- cs.AI Aligning Language Models with General Preferences: An Empirical Study of Direct Nash Optimization
- cs.AI Suppressing Misaligned Behaviors in Language Models, Durably
- cs.AI Adversarial Speculative Reasoning Improves Long Horizon Agents
- cs.AI Weight-space programming for auditing clinical decision networks
- cs.AI MedPhysBench: A Benchmark for Evaluating Language Models on Medical Physics Calculations and Error Detection
- cs.AI A simple adaptive accelerated optimistic dual averaging method for unconstrained convex optimization in $p$-norm geometry
- cs.AI PyRUA-DataFarm: LLM Agent as a Cost-Efficient Teleoperator for Real-Ready Robot Data
- cs.AI HARP: Training-Free Dual-Profile Agentic Communication for LLM-Based Recommendation
- cs.AI You don't need long context
- cs.AI LoopVQ: Sub 1-bit KV Cache Quantization for Looped Transformers
- cs.AI Personas Are Not People: Identity as State, Not Attribute
- cs.AI False Reassurance in LLM Severe-Weather Guidance: A Replayable Evaluation Against Real NWS Alerts
- cs.AI Knowingly Wrong: Training LLMs Against Their Own Beliefs can Induce Misalignment
- cs.AI Beyond Accuracy: What Language-Subspace Interventions Capture, Remove, and Improve
- cs.AI The Geometry of Ignorance: Detecting Suppressed Knowledge in LLMs through Activation Fragmentation
- cs.AI JHiMA-WebRL: A Joint Hierarchical Multi-Agent Reinforcement Learning Framework for Long-Horizon Web Tasks
- cs.AI Going Deeper With Memories
- cs.AI Fairness Metrics based on the AUC in Clinical Predictive AI
- cs.LG PolicyIK: High-Precision Inverse Kinematics for Robot Arms with Reinforcement Learning
- cs.LG GUILD: Guided Iterative Data Collection for Imitation Learning
- cs.LG Reading Why: Born-Measured Attention with Exact Word Evidence on Light and Qubits
- cs.LG Agnostic Learning by Imitating
- cs.LG A Circuit for In-Context Learning
- cs.LG CATF: A Manager-Worker Framework for Context-Aware Time-Series Forecasting
- cs.LG $Φ$asmaNet: A Staged Spectral Token-Refinement Architecture for Fast Gravitational-Wave Signal-to-Noise Ratio Prediction
- cs.LG Universal Hardness From Masking Channels
- cs.LG GSBR: Geometric Support Behavior Regularization for Online RL
- cs.LG Conformal Prediction at Performative Equilibrium: The Price of Performativity
- cs.LG De Rham cohomological architectures for PINNs and PDEs with topological holes
- cs.LG iGC: Antibody Optimization via Reinforcement Learning over a Learned Somatic Hypermutation Prior
- cs.LG Stabilizing On-Policy Targeted Guidance in Agentic Reinforcement Learning
- cs.LG Gaussian Partitioning for Reliable Learning of Multi-Index Models
- cs.LG NovoCRF: Globally Normalized De Novo Peptide Sequencing on a Prefix-Mass Lattice
- cs.CL MDSKILL: Learning to Internalize Skill Memory through Multi-View Distillation for Efficient Dynamic Agent Self-Evolution
- cs.CL ARLoop: Fully-Open Pre-trained Looped Transformers with Attention Residuals
- cs.CL What Does the Model See? Representation Validity in Multimodal LLM-Based Video Analysis
- cs.CL When Self-Verification Hurts: Calibrated Answer Repair for Financial Question Answering
- cs.CL GM League Arena: An Interactive Long-Horizon Task Benchmark Based on Professional Sports League Management
- cs.CL The Right Source Is Not Enough: Understanding History Interference in Speech-LLM Memory
- cs.CL ToolSpace: Scaling Tool-Centric Environment Synthesis for Long-Horizon Agents
- cs.CL Whose Effect Is It? How Objectives Inflate Learned Subspace Interventions
- cs.CL Who is the Pronoun Pro? Comparing LLMs and Humans at Number-Disambiguation of 'they'
- cs.CL Shuffled OPSD: When Semantic Consistency Hurts On-Policy Self-Distillation?
- cs.CV Visibility-Aware Reference Generation for Benchmarking Visual 3D Reconstruction in Vegetation-Rich Environments
- cs.CV GSCV++: Lightweight 3D Gaussian Splatting Sequence Compression via Standard Video Codecs
- cs.CV OpenBrain: An Auditable Generated-Label Release for Whole-Brain MRI Parcellation
- cs.CV Wonder 4: A Mixture-of-Experts Recipe for Image Quality Enhancement
- cs.CV Low-Density 3D Point Cloud Classification
- cs.CV Anchored Rank-Adversarial Fine-Tuning Pretrained Visual Place Recognition Models
- cs.CV YANAMI: A Large-Scale Multimodal Dataset and Benchmark for Multiturn Stateful Interaction
- cs.CV Möbius: A Unified Motion-Language Model Bridging Understanding and Generation
- cs.CV Route Anything: A Study of Model and Policy Routing for General-Purpose AI Systems
- cs.CV Diagnosis by Composition: Detecting and Localizing Unseen Diseases by Composing Learned Radiological Findings
- cs.CV PLANTEUR: A Crop-Type Segmentation Benchmark for Tropical Regions
- cs.CV Diffusion Plug-in Priors for Self-Supervised Ultrasound Video Restoration
- cs.CV TAQ-MOTR: Tracking-Aware Detection Query Formation for End-to-End Multiple Object Tracking
- cs.CV Robust Reconstruction of Power Lines from Incomplete Mobile and Aerial LiDAR Data
- cs.CV Factorized MLLM Embeddings for Composed Image Retrieval
- cs.CV From Natural Images to Medical Scans: Transferability of Diffusion Models for Medical Image Generation
- cs.CV PadLoc: Training-Free Floor-Plan Localization by Sparse Geometric Matching
- cs.CV Perturb, Repair, Verify: Self-Play Vision-Language Verifiers for Compositional Understanding
- cs.CL They Know Where It Is, But Not What It Says: A Bilingual Hallucination Benchmark of Hadith Recall in Open LLMs
- cs.CL From Language to Country to Valence: A Multilingual Audit of Stereotype Bias Across LLMs
- cs.CL SMILE: A Structured Multi-Agent Clinical Reasoning for Liver Ultrasound Understanding
- cs.CL Re$^3$Rank: Listwise Reranking with Reasoning-Conditioned Retrieval Heads
- cs.CL FindLink: Learning to Assess Scientific Connections from Source-Grounded Premises
- cs.CL PixelBench: Can LLMs Understand and Generate Images Pixel by Pixel?
- cs.CL Tune the Baseline First: Most of Muon's Scaled Weight Decay Gain Is a Tuning Gap, but Its Small Real Advantage Grows With Model Size
- cs.CL Ranked but Not Stopped: Detecting Prompt-Injection Hijacks from Typed Decision Models' Output Probabilities
- cs.CL Same Sentence, Different Construct: Identity-Swapped Rationales Reveal Measurement Non-Invariance in LLM Hate Speech Classifiers
- cs.CL Chain on Demand: Deferred Evidence Organization for Long-Term Conversational Memory
- cs.CL GigaSpeech 3: Scaling Multi-Dialect Arabic Speech Recognition with Weakly Labeled Data
- cs.CL As Expected? Efficient LLM Planning via Expectation-Observation Consistency Verification
- cs.CL Diagnosing Model--Item Interactions in Language Model Benchmarks with LSIRM
- cs.CL InteractionBench: Measuring Timing, Task Accuracy, Whisper and Volume, and Grounding in Real-Time Voice Systems
- cs.CL When Rewards Depend on the Group, GRPO Plays a Tournament
- cs.CL Solver-Like Behavior in Large Language Models as Formalizers
- cs.CL EviCast: Forecasting Retrieval Gains Through Shared Response Calibration
- cs.CL Beyond Accuracy: How Quantization Affects Coding Agents
- cs.CL VisCodeBench: a new benchmark dataset for Text-to-Vis reflecting real-world practice
- cs.CL Which Words a Model Fears: From a Refusal Direction to a Vocabulary-Wide Fear Lexicon
- cs.AI Pixel-UMI: Learning Environment-Aware Whole-Body Manipulation from Sensorized Human Demonstrations
- cs.AI Pixel: A Co-Designed Full-Stack Robot System for Learning General-Purpose Mobile Manipulation
- cs.AI From Natural Language to Exact Cover: A Neuro-Symbolic Approach to Zebra Puzzles
- cs.AI WebHarbor: Docking Real Websites for Evolving GUI Agent Environments
- cs.AI D&A: Decoupling Description and Assessment Enables VLMs as Zero-Shot Robotic Reward Models
- cs.AI Strategist: Evolving Reasoning with Agents
- cs.AI Tracing the Scientific Discovery Capacity of LLMs to Their Training Data
- cs.AI DiabAgent: Toward Agentic Physiological Control
- cs.AI Four Bridges: Measuring Deception Propensity in a Multi-Agent Game
- cs.AI Spacetime Emergence in the Memory of a Learning Agent
- cs.AI Evaluating Uncertainty Calibration in Probabilistic Time Series Foundation Models
- cs.AI FedReady: Bounded Agents for Federation Formation
- cs.AI Sonar-MoE: In-Backward Expert Updates for Memory-Efficient Mixture-of-Experts Training
- cs.AI PlantVLM-Edge: Quantized Vision-Language Models for On-Device Crop Health Diagnosis
- cs.AI MCP Tool Poisoning in Agentic AI Frameworks: A Cross-Framework Measurement Study
- cs.AI Can LLM Agents Be Forward-Deployed Engineers? Requirement Discovery and Delivery for Enterprise Text-to-SQL
- cs.AI An Agentic Framework for Automated Power System Studies and Grid Connection Assessment