← 研踪 PaperTrailarXiv:2611.24697 · v1
Depth Before Breadth: Reverse-to-Forward Curriculum with Advantage Decomposition for LLM Agent
首次提交 2026-11-19 00:25
版本更新 2026-11-19 00:25
雷达观测 2026-11-19 15:04
摘要
Reinforcement learning (RL) has become a common approach for training large language model (LLM) agents on long-horizon tasks. However, learning from sparse terminal rewards makes credit assignment and exploration difficult. We approach these challenges through a policy-induced capability region, the set of states from which the current policy can reliably complete a task. Near its boundary, an agent may need only a few actions to reach a more solvable state, even when the goal remains distant. This reduces long-horizon learning to two coupled subproblems of reaching that state and completing the task from there. We address them with paired advantages and a progressive reverse-to-forward curriculum. The advantages provide separate signals for reaching more solvable states and completing the task from them, while the curriculum expands the capability region backward along successful reference trajectories and then into off-reference states through on-policy exploration. On ALFWorld and WebShop, our method outperforms competitive agentic RL methods at both model scales, achieving success rates of 90.1%/96.1% and 80.2%/82.6% with Qwen2.5-1.5B/7B-Instruct, respectively.
观测到的版本
相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。
v1 · 2026-11-19 00:25 · 最新观测版本
Depth Before Breadth: Reverse-to-Forward Curriculum with Advantage Decomposition for LLM Agent
Reinforcement learning (RL) has become a common approach for training large language model (LLM) agents on long-horizon tasks. However, learning from sparse terminal rewards makes credit assignment and exploration difficult. We approach these challenges through a policy-induced capability region, the set of states from which the current policy can reliably complete a task. Near its boundary, an agent may need only a few actions to reach a more solvable state, even when the goal remains distant. This reduces long-horizon learning to two coupled subproblems of reaching that state and completing the task from there. We address them with paired advantages and a progressive reverse-to-forward curriculum. The advantages provide separate signals for reaching more solvable states and completing the task from them, while the curriculum expands the capability region backward along successful reference trajectories and then into off-reference states through on-policy exploration. On ALFWorld and WebShop, our method outperforms competitive agentic RL methods at both model scales, achieving success rates of 90.1%/96.1% and 80.2%/82.6% with Qwen2.5-1.5B/7B-Instruct, respectively.
查看此版本解读 →查看此版本原文 ↗