2605.26784v1 May 26, 2026 cs.LG

Ratio-Variance Regularized Policy Optimization

Lei Lv
Lei Lv
Citations: 29
h-index: 2
Fuchun Sun
Fuchun Sun
Citations: 352
h-index: 9
Yihan Hu
Yihan Hu
Citations: 24
h-index: 3
Dong Li
Dong Li
Citations: 175
h-index: 5
Jianye Hao
Jianye Hao
Citations: 184
h-index: 5
Yu Luo
Yu Luo
Citations: 752
h-index: 8
Shuo Han
Shuo Han
Citations: 4
h-index: 1
Huaping Liu
Huaping Liu
Citations: 507
h-index: 4

Standard on-policy reinforcement learning relies on heuristic clipping to enforce trust regions, but this mechanism imposes a severe cost by indiscriminately truncating high-return yet high-divergence updates. We demonstrate that explicitly constraining the policy ratio variance provides a principled local approximation to trust-region constraints, eliminating the need for binary hard clipping. By acting as a distributional ``soft brake'', this approach preserves critical gradient signals from novel discoveries while naturally down-weighting and enabling the reuse of stale, off-policy data. We introduce ${\bf R}^2{\bf VPO}$ (Ratio-Variance Regularized Policy Optimization), which implements this constraint via a primal-dual optimization framework. Extensive evaluations across $7$ LLM scales, spanning both fast and slow reasoning paradigms, and $10$ robotic control tasks demonstrate the generality of the proposed approach. R$^2$VPO achieves substantial performance gains on mathematical reasoning benchmarks, with particularly pronounced improvements on smaller models, while significantly improving sample efficiency. Furthermore, it consistently outperforms PPO baselines in continuous control domains, particularly in sparse-reward and dynamic environments. Together, these findings establish ratio-variance regularization as a principled foundation for stable and data-efficient policy optimization.

0 Citations
0 Influential
4.5 Altmetric
22.5 Score
Original PDF

No Analysis Report Yet

This paper hasn't been analyzed by Gemini yet.

Log in to request an AI analysis.

댓글

댓글을 작성하려면 로그인하세요.

아직 댓글이 없습니다. 첫 번째 댓글을 남겨보세요!