Safety
Bellman Policy Optimization
arXiv:2609.15987v1 Announce Type: cross Abstract: Reinforcement learning with verifiable rewards (RLVR) improves the reasoning capabilities of large language models (LLMs). We introduce Bellman Policy
arXiv:2609.15987v1 Announce Type: cross Abstract: Reinforcement learning with verifiable rewards (RLVR) improves the reasoning capabilities of large language models (LLMs). We introduce Bellman Policy Optimization (BPO), a critic-free method derived from Policy Mirror Descent (PMD). For autoregressive generation with terminal rewards, BPO uses the Bellman equations to reformulate PMD as a trajectory-level objective. The reformulation avoids estimating state values at intermediate states. We prove that it has the same unique optimal solution as the original PMD objective. We derive the practical BPO loss by approximating this objective. Its mismatch-correction weight is a smoothed ratio of complementary token probabilities. Experiments on mathematical reasoning benchmarks demonstrate the effectiveness of BPO.
Related
- Adapt to Thrive! Adaptive Power-Mean Policy Optimization for Improved LLM Reasoning
- Perception-Aware Policy Optimization for Multimodal Reasoning
- When Self-Belief Misleads: Active Label Acquisition for Reinforcement Learning with Verifiable Rewards
- ResRL: Boosting LLM Reasoning via Negative Sample Projection Residual Reinforcement Learning
- GRPO-VPS: Enhancing Group Relative Policy Optimization with Verifiable Process Supervision for Effective Reasoning
Source: arXiv cs.CL | 2026-09-15