Papers

Filtered to skill-based reinforcement learning · clear filter

Browse by term

continual learning 64reinforcement learning 34large language models 17benchmarking 12vision-language models 10generative models 8language models 8video generation 7multimodal models 6natural language processing 6robotics 6world models 6benchmarks 5diffusion models 5on-policy distillation 5policy optimization 5scalability 5self-distillation 5vision-language-action models 5autoregressive models 4computer vision 4diffusion transformers 4LLMs 4multimodal large language models 4verifiable rewards 4attention mechanisms 3embodied intelligence 3image editing 3long-term memory 3multimodal learning 3

Matching papers

Skill Self-Play: Pushing the Frontier of LLM Capability with Co-Evolving Skills

34 upvotes · 24 JUL 2026 · Siyuan Huang, Pengyu Cheng, Haotian Liu et al.

This paper develops a new framework called Skill Self-Play that helps large language models (LLMs) improve their capabilities by co-evolving skills that balance task diversity and verification reliability. Practitioners might care because this approach can lead to significant performance gains for LLMs.

continual learningskill-based reinforcement learningco-evolutionary methodstask generationmodel improvement