arXiv AI

MileGPO: Milestone Inference with Local Evidence for Graph-Based Policy Optimization of Long-Horizon LLM Agents

arXiv:2608. 19803v1 Announce Type: cross Abstract: Credit assignment is challenging in long-horizon agentic reinforcement learning, where supervision often comes only from final rewards.

arXiv Machine Learning
Sep 4

TIGPO: Temporal Instance-Graph Policy Optimization for Long-Horizon LLM Agents

TIGPO (Temporal Instance-Graph Policy Optimization) extends graph-based credit assignment for long-horizon LLM agents by maintaining a persistent transition graph per task across policy updates. It allocates rollout budgets to both new exploration and revisiting past tasks, pairing current rollouts with earlier ones to create cross‑temporal references that stabilize advantage estimation. Experiments on ALFWorld and WebShop show TIGPO consistently outperforms previous group‑based and graph‑based policy optimization methods.

By Jinwei Gan
arXiv AI
Jun 2

Beyond Trajectory-Level Attribution: Graph-Based Credit Assignment for Agentic Reinforcement Learning

arXiv:2605. 26684v2 Announce Type: replace-cross Abstract: Group-based reinforcement learning (RL) methods have achieved remarkable success in improving the performance of large language models (LLMs) and have been rapidly extended to agentic tasks.

By Xin Cheng, Shuo He, Lang Feng, HaiYang Xu, Ming Yan, Lei Feng, Bo An
arXiv AI
Sep 3

PGPO: Potential-Guided Policy Optimization for Multi-Turn Agentic Tasks

The paper introduces Potential-Guided Policy Optimization (PGPO), a method for multi-turn agentic tasks that improves credit assignment by estimating empirical state potentials from anchor-state-group return statistics. PGPO derives action advantages from potential differences between adjacent states, enabling cross-trajectory credit propagation and finer-grained step-level credit assignment, especially within failed trajectories. Experiments on ALFWorld and WebShop demonstrate strong performance compared to recent group-based reinforcement learning methods, with negligible training overhead.

By Yuyao Zheng, Haipeng Sun, Junwei Bao, Lemao Liu, Hongfei Jiang, Yang Song, Dejing Dou
Hugging Face Trending Papers
Jun 24

Semantic Consistency Policy Optimization for Reinforcement Learning of LLM Agents

Group-based reinforcement learning effectively post-trains LLM agents for long-horizon, sparse-reward tasks by deriving step-level credit from trajectory outcomes. However, this ties a step's credit to its rollout's final outcome: semantically near-identical intermediate steps receive opposite credit depending on whether their trajectory eventually succeeded or failed.

arXiv Machine Learning
Aug 31

VICT: Verifier-Instrumented Credit Tracing for Long-Horizon LLM Agent Reinforcement Learning

The paper introduces VICT, a method that leverages the internal structure of verifiable tasks to perform fine‑grained credit assignment for long‑horizon LLM agents. VICT exposes executable or evidence‑backed atoms from a task’s terminal verifier and traces them back to actions via dependency‑valid proof edges, redistributing advantage only along these edges. This approach improves performance on ALFWorld and WebShop compared to outcome‑only training and matches recent fine‑grained credit methods without requiring additional critics, labels, or inference‑time verifier access.

By Pengcheng Li, Zhengyang Zhang, Dongxu Zhang, Sui Huang, Shaohua Ma
arXiv Machine Learning
Jun 25

Neglected Free Lunch from Post-training: Progress Advantage for LLM Agents

arXiv:2606. 26080v1 Announce Type: new Abstract: Process reward models enable fine-grained, step-level evaluation of LLMs, yet building them for agentic settings remains prohibitively difficult: long-horizon interactions, irreversible actions, and stochastic environment feedback make both human annotation and Monte Carlo estimation infeasible at scale.

By Changdae Oh, Wendi Li, Seongheon Park, Samuel Yeh, Tanwi Mallick, Sharon Li
arXiv AI
Jun 8

StainFlow: Entity-Stain Tracking and Evidence Linking for Process Rewards in GUI Agents

arXiv:2606. 07027v1 Announce Type: new Abstract: Reinforcement Learning (RL) has become a promising approach for improving GUI Agents in long-horizon, stochastic digital environments, but trajectory-level success feedback is too sparse to provide reliable credit assignment for intermediate exploration steps.

By Haojie Hao, Longkun Hao, Yihang Lou, Yan Bai, Zhenyang Li, Zhichao Yang, Dongshuo Huang, Hongyu Lin, Lanqing Hong, Jiakai Wang, Xianglong Liu