Reinforcement Learning

Reasoning Arena: Trace Tournaments When Verifiable Rewards Fall Short
Trace Tournaments When Verifiable Rewards Fall Short.
Han Zhou, Adam Yang, Laurence Aitchison, Anna Korhonen, Albert Jiang
Agentic Policy Optimization via Instruction-Policy Co-Evolution
Dynamic instruction optimization in reinforcement learning for agents.
Han Zhou, Xingchen Wan, Ivan Vulić, Anna Korhonen