每日论文 Daily Papers

来自 Hugging Face 与研究社区 · 中英对照

Daily Weekly Monthly

‹ 2026-09-07 ~ 2026-09-13 ›

跳转日期： 11-01 39 10-28 35 10-27 31 10-24 50 10-23 52 10-22 60 10-21 64 10-20 64 10-17 83 10-16 74 10-15 99 10-14 22 10-13 58 10-11 4 10-10 68 10-09 61 10-07 72 10-03 82 10-02 84 10-01 86 09-30 95 09-29 31 09-27 22 09-25 29 09-24 23 09-23 30 09-22 30 09-20 1 09-19 23 09-18 23

Compile by Training: Turning Natural-Language Specifications into Local Neural Functions

▲ 314 💬 3 ★ 4

Terminal-Universe: Turning Agent Trajectories into Scalable Terminal Environments

▲ 268 💬 1

LLaDA-Image: Building Strong Image Generators with Fully Open Training Recipes

▲ 226 💬 3

Random Attention: Rethinking KV Cache Eviction for Efficient Reasoning

▲ 161 💬 4 ★ 31

Knowing When Not to Reuse: Conditional Experience Transfer in Autonomous LLM Post-Training

▲ 149 💬 2

NCP-ArchPreview Technical Report: Moving towards Latent Space Language Models through Next Concept Prediction

▲ 146 💬 1

SenseNova-U1.5: Towards Native Unified Visual Intelligence

▲ 120 💬 1

RoboTok: An Internet-Scale Data Engine for Human Demonstration Retrieval and Dexterous Manipulation Learning

▲ 112 💬 2 ★ 21

LatentPress: Context Compression Beyond Text and Vision

▲ 108 💬 2 ★ 3

Bilevel Coordinated Reflection: A Game-Theoretic Approach to Multi-Agent LLM Systems

▲ 91 💬 1

Show-Harness: Just a VLM Agent Can Play Robots

▲ 85 💬 2 ★ 125

Rethinking On-Policy Distillation of Large Language Models II: One Training Example

▲ 75 💬 2 ★ 37

Why Gated DeltaNet Survives 4-Bit Quantization: NVFP4 W4A4 for the Recurrent Half of a Hybrid 27B LLM

▲ 73 💬 2

Puffin-World: Scaling a Unified Multimodal Model with Native 3D World States

▲ 66 💬 3

Unlocking Lossless Speedups in LLMs via Discrete Diffusion

▲ 64 💬 6 ★ 44

Programmable World Model

▲ 62 ★ 100

AgentGrad: Intervention-guided Prompt Optimization for Multi Agent Systems

▲ 56 💬 1

SpatialBlock: Enhancing Spatial Intelligence in LVLMs via Synthetic Block-Stacking Problem

▲ 52 💬 1

Iris: Climbing to the Search Frontier

▲ 50 💬 3 ★ 32

FlowBalance: Verifier-Grounded Self-Improvement from On-Policy Reasoning Experience

▲ 50 💬 1 ★ 2

Scal3R: Learning Efficient Multi-Relative Pose Query for Scalable Online 3D Reconstruction

▲ 46 💬 2

Editable Visual Design

▲ 41 💬 2 ★ 468

EvoSafeHarness: Evolving Model- and Domain-Specific Harnesses for Securing Agents

▲ 36 💬 1 ★ 1

Motion-Omni: End-to-End Joint Speech and Full-Body Motion for Spoken Dialogue

▲ 35 💬 1 ★ 5

The Missing Temporal Link: Temporal Context Routing for Script-Driven Audio-Video Generation

▲ 33 💬 2

Beyond Retrieval: Progressive Latent Memory Evolution for Streaming Video Understanding

▲ 29 💬 2

WearableQA: A Benchmark for Health Reasoning over Real-World Wearable Data

▲ 28 💬 1 ★ 9

Last Translation Benchmark

▲ 27 💬 2 ★ 50

RealSWE: A Compositional Evaluation of Coding Agents under Realistic User Requests

▲ 25 💬 4

The Attention Triangle in Audio-Video Models

▲ 25 💬 2 ★ 1

CORE: Improving Compositional Reasoning in MLLM Embedding via Reranker Distillation

▲ 24 💬 2

DRACO: Fine-Grained Credit Assignment with Dynamic Rubrics for Long-Horizon Agent Training

▲ 23 💬 5 ★ 2

WorldReward: Reward Modeling for Camera-Conditioned World Models

▲ 23 💬 2

PACE: Towards Surfacing Hidden Conflicts in User Requests

▲ 22 💬 2

WorldSculpt: Generating Compositional Worlds from Grounded Videos

▲ 21 ★ 32

ENEAS: Embedding-guided Neural Ensemble for Adaptive Segmentation

▲ 20 💬 2 ★ 41

Mi-Ripple: Restoring Images Degraded by Iterative AI Editing

▲ 20 💬 2 ★ 32

FlashRender: Few-Step Generative Rendering via Camera-Controlled Video MeanFlow

▲ 19 💬 2 ★ 9

Using Grounded Theory for Agent Behavior Analysis at Scale

▲ 18 💬 2

SWE-Bench Pro Verified: A Reliable Benchmark for Software Engineering Agents

▲ 18 💬 3

Environment Evolution for Terminal Agents

▲ 17 💬 1

Principia: Relational Physics Tests for Video Models

▲ 16 💬 1

Enoki: Efficient Multi-Level Hallucination Detection

▲ 16 💬 1 ★ 11

SAEScientist-Bench: Can AI Agents Conduct Autonomous SAE Interpretability Research?

▲ 16 💬 1 ★ 14

Select, Compress, Reinvest: A Controlled Study of Visual-Token Allocation in Long-Video MLLMs

▲ 15 💬 3 ★ 1

Ask Before You Optimize: Dynamic Pre-Formulation Clarification for Interactive Optimization

▲ 15 💬 2 ★ 4

Scores Alone Do Not Prove Discovery: The Discovery Certification Protocol for Auditing AI Research Agents

▲ 15 💬 1

Puppeteer: Object-Grounded Posture-Aware Co-Speech Gesture Generation

▲ 15 💬 2

Let Confidence Change, Not the Prediction: Prediction-Preserving Repair for Post-hoc Calibration

▲ 13 💬 2

Percolation Dynamics in Optimization : Variance Cascades and Discrete Scale Invariance

▲ 13 💬 2

QCell: Recombining and Aligning Cell Queries for Overlapping Instance Segmentation

▲ 13 💬 5 ★ 9

Don't Drop Dropout: Optimizing Layer Sparsity for Efficient LLM Training and Inference

▲ 13 💬 1

VeriPhy: Agentic Physical Reasoning for World Model Evaluation and Refinement

▲ 12 💬 2

X-AuT: Progressive Audio-Encoder Compression for Speech LLMs with Cross-Scale Distillation

▲ 12 💬 1 ★ 7

SyncWorld: Visual Calibration Enables World Models as Zero-Shot Simulators

▲ 11 💬 1 ★ 15

Beneath the Surface of Chains-of-Thought: A Mechanistic Interpretation of Reasoning Operations in LLMs

▲ 10 💬 1

EmbodiedSkills: A Unified Framework for Orchestrating, Training, and Deploying VLA Agents

▲ 10 💬 1 ★ 26

One Symptom, Three Levers: A Critical Review of On-Policy Self-Distillation

▲ 10 💬 1

A Common Measure of Communication for Speech Brain-Computer Interfaces

▲ 9 💬 2 ★ 3

Locked at the Entrance, Open Inside: Where RLVR Narrows the Solution Space

▲ 9 💬 2 ★ 1

RISE: Recursive Improvement via Self-Extrapolating Policy Distillation

▲ 9

UniMate: One Unified Model to Animate Diverse Skeletons

▲ 9 💬 1 ★ 49

MaxKernel: Agentic Kernel Generation for TPUs

▲ 9 ★ 62

DianShi-RxnDB: A Large-Scale, Fine-Grained Organic Reaction Data Platform Built via a Fully Automated Pipeline for Researchers and AI Agents

▲ 9 💬 1

Memory as Plans: World-Action Modeling with Memory-Grounded Planning

▲ 9 💬 1

Causal Foundation Models

▲ 8 💬 1 ★ 10

Verify Before You Distill: Prompt-Level Teacher Gating for On-Policy Distillation

▲ 8 💬 1

Revisiting Complete Reasoning Traces for Post-Training

▲ 8 💬 1 ★ 2

Φ-Bench: Can Large Language Models Engineer the Infrastructure That Powers Them?

▲ 7 ★ 55

Train Smarter, Not Harder: Switching Signal-Guided Training in Active Learning

▲ 7 💬 1

AgenticGen: Reward-Guided Agentic Video Generation for Advertising

▲ 7

One Editor, Many Edits: A Unified Training-Free Framework for Diverse Video Editing

▲ 6 💬 1

Recursive Code World Models: Building Complex Worlds through Recursive Scene Programs

▲ 6

TempCloze: Can Video-LLMs Identify the Missing Middle?

▲ 6 💬 1 ★ 5

Dr. Claw: An AI Scientist Workspace for Vibe Research

▲ 5 💬 1 ★ 1058

Group Adaptive Clipping Policy Optimization

▲ 5 💬 1 ★ 5

Unifying Conformal Language Tasks with In-Context Ensembles

▲ 5 💬 1 ★ 2

StochBench: A Domain-Specific Benchmark for Stochastic Processes in Lean

▲ 5 💬 1

FreeFlow: A Bias-free Hierarchical Transformer for Optical Flow Estimation

▲ 5 💬 1 ★ 2

World in World: Explore the World with World Models

▲ 5

An Open Recipe for IMO Gold: Training Nemotron for Olympiad Mathematics

▲ 5

τ^τ-Bench: An Environment for End-To-End, Realistic Agent Construction

▲ 4 💬 1 ★ 2

When Models Edit Too Much: On the Fidelity of Minimal Code Edits

▲ 4 💬 1 ★ 19

Training-Free Speech-Centric Omni Understanding with Frozen VLMs

▲ 4 💬 1 ★ 9

Safety for Whom? Boundary-Aware Self-Distillation for Controlled LLM Safety Refusal

▲ 4 💬 1

Why Is Video Still So Expensive? A Survey of Inference-Efficiency Mechanisms in Video and Audiovisual LLMs

▲ 4 💬 1 ★ 2

Co-Evolving Harnesses and Models: On-Policy Correction Helps Weaker Models Catch Up Where Imitation Fails

▲ 4 💬 1

OracleZoom: On-Policy Self-Distillation Inspired Reference-Constrained Recursive Image Super Resolution

▲ 4 💬 1

Diffs vs. Whole Files: An Empirical Comparison of Iterative Edit-Based and Direct Generation for Flutter/Dart Code Models

▲ 4 💬 1 ★ 1

A Three-Layer Caching Architecture for Low-Latency LLM Web Search on Commodity CPU Hardware

▲ 4 💬 1 ★ 22

Negative Self-Distillation: Learning to Reason by Avoiding Flaws

▲ 4 💬 1 ★ 4

CARDEA: Auditable Reasoning Grounded in Spatial Evidence for End-to-End Coronary Angiography Interpretation

▲ 4 💬 1

When Quantization Breaks Memory: Recurrent-State Write-Back in Low-Precision Temporal Inference

▲ 3 💬 1

ShallowStream: Index Shallow then Answer Deep for Streaming Video Understanding

▲ 3 💬 1 ★ 1

To See a World in a Living Context: Unified Indoor-Outdoor Urban World Generation

▲ 3 💬 1

What Else Needs Fixing? Exploring Cost-Effective Test-Time Compute for Revision Propagation in Artifacts Generated Through Conversation

▲ 3 💬 1

RESCUE-BENCH: Towards Relation-Aware Multi-Party Emotional Support Conversation Systems

▲ 3 💬 1 ★ 1

SchemeArena: Factorized Stress Testing of Scheming in LLM Agents

▲ 3 💬 1

PARSER: Read in Parallel, Reason in Depth for Long-Context LLM Agents

▲ 3 💬 1 ★ 2

MetroLLM-Bench: Evaluating Language Models as Transit Kiosk Runtimes

▲ 3 💬 2

HyQuant: Hybrid-Precision Quantization for LLM Attention

▲ 3 💬 1 ★ 8

AdaptVPR: Route-Aware Hard Positive Generation for Robust Visual Place Recognition

▲ 2 💬 1 ★ 2

The Semantic Bottleneck: Leveraging Semantic Representations for Non-Invasive Speech Decoding

▲ 2 💬 1

Difficulty-Adaptive Tree-Structured Policy Optimization for Expanding Reasoning Coverage in RLVR

▲ 2 💬 1

From Reweighting to Rewriting: Unlocking the Intervention Effects of Influential Samples in Training Data Attribution

▲ 2 💬 1

DRG-MAPPO: Hierarchical Dynamic Role-Graph Multi-Agent Reinforcement Learning for Cooperative Air Combat

▲ 2 💬 1

UniH^3: Unifying Hierarchical Homogeneity and Heterogeneity for All-in-One Medical Image Restoration

▲ 2 💬 1 ★ 5

Generative Late-Interaction Embeddings For Visual Document Retrieval

▲ 2 💬 1 ★ 2

Knowing What Not to Answer: Selective Non-Compliance in Vision-Language Models

▲ 1 💬 1 ★ 1

Refuse without Refusal: A Structural Analysis of Safety-Tuning Responses for Reducing False Refusals in Language Models

▲ 1 💬 1 ★ 1

HarvestBench: Measuring Whether LLM Agents Will Pay to Avoid Killing Animals

▲ 1 💬 1

The 2026 PNPL Competition: Word Classification and Efficient Cross-Subject Generalisation in LibriBrain100

▲ 1 💬 1

Reference-Based Bias Detection in LLMs via Relative Representations of Hidden States

▲ 1

PlannerForge: LLM Agents for Scenario-Based Testing of Motion Planners in Autonomous Driving

▲ 1 💬 1 ★ 1

DF26: We Cannot Tell Fake From Real Anymore

▲ 1 💬 1

Beyond Solver Verdicts: Generative Reward Models for Autoformalization

▲ 1 💬 1

Building Multilingual Bridges: Data Mixing as the Pillar of Generalization for In-Language Reasoning

▲ 1 💬 1

Adaptive Bridge: A Proxy-Based Decoupling Layer for Mitigating DDS Backpressure in ROS 2

▲ 1 💬 1

The Price of Sparsity: Sufficient Conditions for Sparse Recovery using Sparse and Sparsified Measurements

▲ 0 💬 1

IdeaAMBIG: Benchmarking Implementation-Critical Gaps in Research-Idea Specifications

▲ 0 💬 1

ActReview: Rebuttal-Guided Training Data and Rubric Rewards for Actionable Peer Review Generation

▲ 0 💬 1