… The following papers were recommended by the Semantic Scholar API Agentic Reinforcement Learning with Self-Distilled Reward Shaping 2026 EviSD: Evidence-Conditioned Self-Distillation for Search-Augmented Agents 2026 OPID: On-Policy Skill Distillation for Agentic Reinforcement Learning 2026 ATOD: An… …
… We introduce Policy Optimization with Internal State Value Estimation , which obtains a baseline at negligible cost by using the policy model's internal signals already computed during the policy forward pass. …
… Code and reproducible scripts are open-sourced in the repo. the core idea that really sticks is target decoupling: keep multi-timescale predictions on the critic for auxiliary representation learning, while the actor updates are driven only by long-horizon advantages. this separation seems to block… …
… In heterogeneous training systems , the total importance ratio should ideally be decomposed into two semantically distinct factors: a training--inference discrepancy term that aligns inference-side and training-side distributions at the same behavior-policy version, and a policy-staleness term that… …
… We develop a statistical theory for offline policy optimization from such outcome-level supervision. …
… In closed-loop simulation , the driving policy model actively interacts with the environment, where its actions dynamically update the simulator state and directly influence the next set of generated sensor observations. …
… Intervention-aware credit assignment further restricts outcome labels to segments executed by the current policy, preventing supervision from leaking across intervention boundaries. …
… On long-CoT mathematical reasoning benchmarks, CSCR consistently outperforms GRPO baseline with the same number of policy updates. …
… Task decompositions are often manually defined according to human semantics, or ignored entirely by treating a whole dataset/domain as “one task.” In both cases, fine-grained structure can be lost: training data is heterogeneous, imbalanced, and composed of problem types whose learning dynamics can… …
… Standard policy gradients pick up on this signal immediately, and within a few hundred updates the model converges to generating identical maximum-length paths for every user, with near-zero diversity. …