Showing cs.LGShow all
2 papers · 1 filter
cs.LG2026
Unveiling Implicit Advantage Symmetry: Why GRPO Struggles with Exploration and Difficulty Adaptation
Zhiqi Yu, Zhangquan Chen, Mengting Liu +2
Reinforcement Learning with Verifiable Rewards (RLVR), particularly GRPO, has become the standard for eliciting LLM reasoning. However, its efficiency in exploration and difficulty…
cs.LG2025
Efficient Skill Discovery via Regret-Aware Optimization
He Zhang, Ming Zhou, Shaopeng Zhai +2
Unsupervised skill discovery aims to learn diverse and distinguishable behaviors in open-ended reinforcement learning. For existing methods, they focus on improving diversity throu…