59 papers
Recursive Synthesis for Long-Horizon Terminal Tasks
Zhongzhi Li, Yucheng Shi, Zongxia Li +8
High-quality long-horizon training data for terminal agents is expensive to produce, often costing hundreds to thousands of dollars per task, because each task must keep the instru…
Stale but Stable: Staleness-Adaptive Trust Regions for Stabilizing Asynchronous Reinforcement Learning
Junyao Yang, Yucheng Shi, Zongxia Li +6
Asynchronous reinforcement learning improves throughput by decoupling rollout generation from optimization, but the resulting staleness is an inevitable byproduct, compounded joint…
Harness Handbook: Making Evolving Agent Harnesses Readable,Navigable, and Editable
Ruhan Wang, Yucheng Shi, Zongxia Li +7
The paper presents the Harness Handbook, a tool that automatically creates a behavior‑centric view of AI agent harness code using static analysis and LLM assistance, enabling devel…
DeepSearch-World: Self-Distillation for Deep Search Agents in a Verifiable Environment
Xinyu Geng, Xuanhua He, Sixiang Chen +7
The paper introduces DeepSearch-World, a deterministic, verifiable web environment, and DeepSearch-Evolve, a self‑distillation framework that lets web search agents improve from th…
Hierarchical Sparse Attention Done Right: Toward Infinite Context Modeling
Xiang Hu, Xinyu Wei, Hao Gu +10
Scaling modern large language models (LLMs) to long contexts is limited by the quadratic computation cost, and poor length extrapolation of dense attention. Chunk-wise sparse atten…
Dual-Uncertainty Guided Policy Learning for Multimodal Reasoning
Rui Liu, Dian Yu, Tong Zheng +8
Reinforcement learning with verifiable rewards (RLVR) has advanced reasoning capabilities in multimodal large language models. However, existing methods typically treat visual inpu…