6 papers
When Sparse Reward Meets Dense Distillation: Training Dynamics of On-Policy Distillation
Xinke Jiang, Tao Feng, Zhibang Yang +4
Reinforcement learning with verifiable rewards provides a sparse post-training signal: a single binary outcome evaluates the entire rollout, and every token receives the same seque…
GenMem: Generative Symbolic Memory for Self-Evolving Harness
Xinke Jiang, Tao Feng, Weixuan Xu +7
Long-term memory supports the self-evolution of LLM agents by retaining experience and skills across tasks and enabling their retrieval, reuse, and revision in subsequent long-hori…
AgenticRag-R1: Agentic Reinforcement Learning with Stack Memory for Multi-Step Reasoning, Retrieval and Memorizing
Xinke Jiang, Yue Fang, Zhibang Yang +12
Retrieval-Augmented Generation (RAG) improves the factuality of large language models (LLMs), yet existing RAG systems often struggle with complex, multi-step reasoning that requir…
SERL-SQL: Selective Hindsight Distillation for Text-to-SQL Reinforcement Agentic Learning
Tao Liu, Tao Feng, Xiangheng Li +9
Recent Text-to-SQL systems increasingly rely on multi-turn interaction, execution feedback, and reinforcement learning. However, most existing methods use execution correctness onl…
Fishing Out Free Riders: Shapley-Based Reward Attribution for Parallel Reasoning via Reinforcement Learning
Wentao Zhang, Haoyu Zhang, Xinke Jiang +7
Large Language Models (LLMs) excel at multi-step reasoning, yet current parallel reasoning approaches often fail to distinguish the contributions of individual reasoning paths. Man…
KbSD: Knowledge Boundary aware Self-Distillation for Behavioral Calibration in Agentic Search
Tao Feng, Xinke Jiang, Chao Wu
Agentic search equips large language models with dynamic retrieval abilities, but existing reinforcement learning methods remain limited by reward sparsity in knowledge boundary ca…