Hai-Xin Wang

This work introduces ACLArena, a framework for comprehensively studying, analyzing, and evaluating Agent Continual Learning, and proposes a new ACL recipe that combines offline replay over high-quality trajectories with a routed network of multiple LoRA experts each specialized via RL, substantially improving the agent's ability to learn across multiple domains.

A novel neural network framework achieves high accuracy in solving PDEs on unbounded domains by intelligently decomposing spatial regions into specialized subnetworks.

HarnessBridge can outperform specialized harnesses while cutting token usage and trajectory length, revolutionizing LLM agent interactions.