雷达日报 · 2026-09-10

今日 65 条信号 · SI 正常 · 精选 4

🎯 今日精选

1. Procedural Graphs: Self-Evolving Execution Structures for LLM Agents

2609.09153 · ? · HF 🔥23 · #untagged 核心 idea：Large language models are increasingly deployed as agents that plan over long horizons and act through external tools. arXiv

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

2609.06703 · ? · HF 🔥4 · #untagged 核心 idea：High-quality structured organic reaction data are essential for developing artificial intelligence for chemistry (AI4Ch… arXiv

3. EVOHARNESSBENCH: Can Your Agents Keep Pace with an Evolving Harness?

2609.04280 · ? · HF 🔥4 · #untagged 核心 idea：Modern LLM-based agents operate through a harness of tools, reusable skills, and specialist agents that shapes what the… arXiv

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

2609.09219 · ? · HF 🔥11 · #untagged 核心 idea：AI research agents combine prior knowledge, public sources, and experimental feedback to produce useful results. arXiv

📋 速览（其余 61 条）

• SAEScientist-Bench: Can AI Agents Conduct Autonomous SAE Interpretability Research? 2609.09113 · 🔥7
• Omni Interaction Agent Technical Report 2609.08977 · 🔥121
• Show-Harness: Just a VLM Agent Can Play Robots 2609.10522 · 🔥31
• Programmable World Model 2609.10540 · 🔥21
• RoboSPA: Can VLA Models Go Beyond Simple Scenes and Short-Horizon Tasks? 2609.05324 · 🔥21
• WearableQA: A Benchmark for Health Reasoning over Real-World Wearable Data 2609.05405 · 🔥16
• Counter-Swarm Doctrine: Containing Coordinated Agent Intrusions 2609.06140 · 🔥3
• AgenticGen: Reward-Guided Agentic Video Generation for Advertising 2609.09187 · 🔥2
• MasterControl Seventeen Every Time 2609.03209 · 🔥1
• Stencil Computation at the Intersection of AI and HPC 2609.10368 · SI 83
• OpenWAM: An Open, Modular Exploration Towards Systematic World-Action Model Pretraining 2609.07398 · 🔥55
• Steering Geometry: Validating Human Value Geometry in LLM Steering Space 2609.06289 · 🔥27
• VDiff-Bench: A Challenging Benchmark for Fine-Grained Image Difference Identification 2609.06245 · 🔥22
• Harnessing CLIP and DINO: An Uncertainty-Aware Cascaded Fusion Network for Generalizable Deepfake Image Detection 2609.07670 · 🔥15
• A*-Thought-V2: Efficient Latent Reasoning via Geometric Dynamics of LLM 2609.07821 · 🔥11
• Φ-Bench: Can Large Language Models Engineer the Infrastructure That Powers Them? 2609.10226 · 🔥3
• Co-Evolving Harnesses and Models: On-Policy Correction Helps Weaker Models Catch Up Where Imitation Fails 2609.09134 · 🔥1
• RESCUE-BENCH: Towards Relation-Aware Multi-Party Emotional Support Conversation Systems 2609.09657
• AXON: A ROS 2 RMW with Shared-Memory/QUIC Transport and QKD/ML-KEM Key Establishment 2609.10024 · SI 65
• Towards a Quantum Erlangen Program 2609.09167 · SI 59
• Academia x Industry: The Role of Fundamentals for Silicon in an AI Native Era 2609.09344 · SI 46
• Measuring Language Transfer in Robot Policies: Adding Greek to a Cosmos3 Vision-Language-Action Policy 2609.07470 · 🔥18
• ReactVAU: A Slow-Fast Decoupled Framework for Streaming Video Anomaly Understanding 2609.07941 · 🔥15
• Recognition-Refusal Misalignment in LLMs: Why Models Answer Structurally Unanswerable Questions 2608.29109 · 🔥15
• Revisiting Complete Reasoning Traces for Post-Training 2609.07103 · 🔥5
• Difficulty-Adaptive Tree-Structured Policy Optimization for Expanding Reasoning Coverage in RLVR 2609.08650
• Where Does the Human End? Creative Agency with Generative AI across Five Years of Chinese Digital Painting 2609.09333 · SI 74
• A Multi-Model Non-Intrusive Reduced-Order Framework for Parametric Erosion Prediction via Kinematic Cross-Moment Compression 2609.09997 · SI 62
• Violet: Enabling Full Virtualization for M-mode RTOS on RISC-V 2609.09833 · SI 51
• An $O(1/T^3)$ algorithm for minimizing convex quadratic functions over the $L_1$ ball 2609.10314 · SI 50
• AuK Technical Report: An Open-Source Foundational Model for Speech Generation and Editing 2609.08936 · 🔥182
• BeaconKV: Key-Value Cache Compression Guided by Beacon Queries for Efficient Large Reasoning Model Inference 2609.04971 · 🔥32
• Reason Through the Latent! Making Latent Visual Reasoning Necessary 2609.06746 · 🔥30
• CosmoH2G: A Hand-to-Gripper Transfer Dataset and Baseline Method for Object Manipulation with Complex Spatial Movements 2609.07498 · 🔥25
• CoVeR: Coverage-Based Token Pruning for Multi-View 3D Reasoning in VLMs 2609.08345 · 🔥20
• TransNormal-2: Geometry-Grounded Rectified Flow with Edge-Aware Decoding for Precise Normal Estimation 2609.06665 · 🔥18
• Cadence: Error-Bounded Lossy Compression of Demand Time Series with a Time-Series Foundation Model 2609.06008 · 🔥16
• SynthGait-19K: A Physically Grounded Synthetic Video Dataset for Gait Parameter Estimation 2609.08108 · 🔥15
• StudyBench: Can Self-Evolution Squeeze Textbooks for Olympiad Capability? 2609.00787 · 🔥12
• Diffs vs. Whole Files: An Empirical Comparison of Iterative Edit-Based and Direct Generation for Flutter/Dart Code Models 2609.05779 · 🔥1
• Puppeteer: Object-Grounded Posture-Aware Co-Speech Gesture Generation 2609.00369 · 🔥1
• Planetary Nebula Central Stars as Tracers of Planetary Nebula-Star Cluster Associations in the Galaxy 2609.09618 · SI 83
• The magnetic field in M16: new results from the JCMT BISTRO survey 2609.09913 · SI 78
• Gradient-Enhanced Proximal Algorithms for Mean Field Planning on Surfaces 2609.10370 · SI 74
• Physics-Informed Multi-Task Surrogate Model for the Martian Nightside Thermosphere 2609.10077 · SI 70
• Chance, Persistent Advantage, and the Generative-AI Era in Open-Source Package Careers 2609.09687 · SI 60
• Exact-Form Regret for Gradient Descent, Mirror Descent and Follow-the-Regularized-Leader 2609.09466 · SI 48
• Field-level prediction of mid-plane stress tensor fields in concrete target penetration: a cross-velocity graph neural operator surrogate 2609.10032 · SI 45
• Eliciting Weak-to-Strong Generalization with On-Policy Reverse Distillation 2609.08798 · 🔥84
• GE-Act 2.0: Pretraining and Scaling a World-Action Model for Robotic Manipulation 2609.05588 · 🔥49
• Marigold V2: Revisiting Diffusion Transformers for Monocular Depth Estimation 2609.08084 · 🔥45
• Mask Forcing: Improving Autoregressive Video Diffusion Distillation via Dual-Noise Masking Rollout 2609.09123 · 🔥43
• TANGO: Humanoid Navigation in Cluttered Environments with a Whole-Body Vision-Language-Action Model 2609.09158 · 🔥19
• What Did I Just Say? Self-Listening for Full-Duplex Speech Models 2609.05592 · 🔥16
• Encoded Early, Used Late: Where Transformers Begin to Act on an Inferred Partner's Expertise 2609.07139 · 🔥15
• SQS: Bayesian DNN Compression through Sparse Quantized Sub-distributions 2510.08999 · 🔥15
• RelightFormer: Feed-forward Generative Transformer for Multiview Object Relighting 2609.07414 · 🔥8
• NOAH: Learning the Full Patient Journey. A Longitudinal Multimodal Time-Aware Model for Representation and Forecasting 2609.09140 · 🔥3
• Graph Machine: Towards Better Pretraining via Edges 2609.02881 · 🔥2
• RenderFormer-V2: Neural Rendering with Heterogeneous Scene Primitives 2609.05738 · 🔥2
• Learning 3D Editing without Paired Supervision via Generative Prior Distillation 2609.04942

🔥 HF 热度榜（Top 5）

1. AuK Technical Report: An Open-Source Foundational Model for Speech Generation and Editing · 🔥182
2. Omni Interaction Agent Technical Report · 🔥121
3. Eliciting Weak-to-Strong Generalization with On-Policy Reverse Distillation · 🔥84
4. OpenWAM: An Open, Modular Exploration Towards Systematic World-Action Model Pretraining · 🔥55
5. GE-Act 2.0: Pretraining and Scaling a World-Action Model for Robotic Manipulation · 🔥49

📡 雷达备注

• ✓ 数据源全部正常

---

下期：明天 21:00（北京时间）