method · first seen 15 Oct 2026
generalized advantage estimation
A method for estimating advantages in policy gradient RL by exponentially weighting multi-step temporal differences.
Mentions per week
PapersCodeDiscussion
Earliest sources
- Terminal-RSI: Recursive Self-Improving Terminal Agents via Recursive Self-Rewrite and Policy-RefinePapers · 15 Oct 2026 · Junyao Yang, Zongxia Li, Yucheng Shi, Zhongzhi Li, Ruhan Wang, Ruiwen Zhou, Hait…
- From Outcomes to Steps: Reliable Success-Aware Critic Learning for Long-Horizon Language AgentsPapers · 20 Oct 2026 · Jiaxi Liu, Shixiong Kai, Zhentao Tang, Yang Li, Yuqi Cui, Mingxuan Yuan, Junchi…
- Learning from Relative Action Credit for Escaping Local MinimaPapers · 27 Oct 2026 · Giovanni Minelli, Giulio Turrisi, Lorenzo Amatucci, Victor Barasuol, Claudio Sem…
Latest sources
- Learning from Relative Action Credit for Escaping Local MinimaPapers · 27 Oct 2026 · Giovanni Minelli, Giulio Turrisi, Lorenzo Amatucci, Victor Barasuol, Claudio Sem…
- From Outcomes to Steps: Reliable Success-Aware Critic Learning for Long-Horizon Language AgentsPapers · 20 Oct 2026 · Jiaxi Liu, Shixiong Kai, Zhentao Tang, Yang Li, Yuqi Cui, Mingxuan Yuan, Junchi…
- Terminal-RSI: Recursive Self-Improving Terminal Agents via Recursive Self-Rewrite and Policy-RefinePapers · 15 Oct 2026 · Junyao Yang, Zongxia Li, Yucheng Shi, Zhongzhi Li, Ruhan Wang, Ruiwen Zhou, Hait…