AI Paper Atlas

Research map / Reinforcement learning / Visual and representation RL

Milestone-Guided Policy Learning for Long-Horizon Language Agents

Zixuan Wang, Yuchen Yan, Hongxing Li, Teng Pan, Dingming Li, Ruiqing Zhang, Weiming Lu, Jun Xiao, Yueting Zhuang, Yongliang Shen

ICML 2026 · Reinforcement learning › Visual and representation RL · approach: In-context and adaptive RL · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use BEACON milestone-anchored credit instead of GRPO on long ALFWorld tasksQwen2.5-1.5B-Instruct · ALFWorld (Long tasks)success rate (%): 53.5 -> 92.9Improved
use BEACON instead of GiGPOQwen2.5-1.5B-Instruct · ScienceWorldsuccess rate (%): 25.8 -> 45.3Improved
add temporal reward shaping and milestone credit to improve sample utilizationGRPO · ALFWorldeffective sample utilization (%): 23.7 -> 82.0Improved
add temporal decay (gamma=0.95) instead of gamma=0 (ablation)BEACON · ALFWorldsuccess rate (%): 81.2 -> 91.4Improved
add temporal decay (gamma=0.95) instead of gamma=0 (ablation)BEACON · WebShopsuccess rate (%): 62.1 -> 75.6Improved
use uniform shaping (gamma=1) instead of decayed shaping (ablation)BEACON · ALFWorldsuccess rate (%): 81.2 (gamma=0) -> 71.8 (gamma=1)Hurt
use detected milestones instead of random partition (ablation)BEACON · ALFWorldsuccess rate (%): 74.2 -> 91.4Improved
drop 50% of milestones (ablation)BEACON · ALFWorldsuccess rate (%): 91.4 -> 82.8Hurt
add segment-level advantage to GRPO's trajectory-level advantage (ablation)BEACON · ALFWorldsuccess rate (%): 72.8 -> 91.4Improved
add trajectory-level advantage to segment-level advantage (ablation)BEACON · ALFWorldsuccess rate (%): 23.4 -> 91.4Improved
add trajectory-level advantage to segment-level advantage (ablation)BEACON · WebShopsuccess rate (%): 67.9 -> 75.6Improved
use BEACON (gamma=0.95) instead of SFT on oracle trajectoriesQwen2.5-1.5B-Instruct · ALFWorldsuccess rate (%): 43 -> 91.4Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy