HF Daily Papers · 2026-10-28

HF Daily Papers 精选（15 篇，按社区票数排序）

  • [189▲] DynaFlow: Learning Free-Form Multi-Agent Orchestration — https://arxiv.org/abs/2610.34467
  • [103▲] Work Agents: Systems, Capabilities, and Optimization — https://arxiv.org/abs/2610.35845
  • [42▲] Spatial Audio Generation and Editing with Executable Plans and Self-Distilling Transfusion — https://arxiv.org/abs/2610.34415
  • [38▲] AV-FABLE: Audio-Visual Manipulation Forensics in Long-Form Videos — https://arxiv.org/abs/2610.33413
  • [33▲] Eat Your Precision: Precision Objective Mitigates Reward Hacking in Rubric-based RL — https://arxiv.org/abs/2610.33512
  • [25▲] Δτ-Bench: Pushing Agents Toward a Moving Capability — https://arxiv.org/abs/2610.34238
  • [24▲] HiP3R: Hierarchical Pose Graphs for Long-Horizon 3D Reconstruction — https://arxiv.org/abs/2610.31635
  • [22▲] Trustworthy Recursive Self-Improvement in the Era of Large Models: Foundation, Risks, Evaluation and Open Challenges — https://arxiv.org/abs/2610.35193
  • [18▲] ProAS: A Progress-Aware Advantage Shaping Framework for Multi-Turn Agentic Reinforcement Learning — https://arxiv.org/abs/2610.18683
  • [17▲] Evolving to Match Capabilities: Improving Harness Self-Evolution via Skill Adaptation — https://arxiv.org/abs/2610.33024
  • [17▲] miniTE: Diagnosing and Bridging the Execution Gap in Native NVFP4 Reinforcement Learning — https://arxiv.org/abs/2610.23928
  • [16▲] S3AC: Stabilized, Selective, and Scalable Actor-Critic for LLM Reasoning — https://arxiv.org/abs/2610.35518
  • [15▲] Build, Don't Play: Self-Improving LLM Research Agents Develop a Hearthstone Battlegrounds AI That Rivals Top Players — https://arxiv.org/abs/2610.33845
  • [14▲] EmbodiedLab: Benchmarking Embodied Agents That Act to Know — https://arxiv.org/abs/2610.35385
  • [13▲] CausalAlign: Causal Alignment of Video Generation Models for Consistent Ego-Motion — https://arxiv.org/abs/2610.29915