函数空间策略优化（FSPO）发布：结合特权适应与函数空间时间整合的强化学习新框架

Function-Space Temporal Self-Distillation for Policy Optimization

要点速览

函数空间策略优化（FSPO）框架通过存储稀疏历史下一词分布作为固定时间目标，结合特权适应与函数空间时间整合，显著改进了 SDPO 方法，无需维持持久教师模型即可实现时间分离。

正文

The paper shows a significant improvement over the recent self-distillation policy methods, SDPO.

This work introduces Function-Space Policy Optimization (FSPO), a single-policy framework that combines current privileged adaptation with function-space temporal consolidation. FSPO stores sparse historical next-token distributions as fixed temporal targets, providing temporal separation without maintaining a persistent teacher model.

Please feel free to discuss the paper if you find it interesting.