← Search

Jian Mu

2 accepted papers

2026

DRIFT: Decoupled Rollouts and Importance-Weighted Fine-Tuning for Efficient Multi-Turn Optimization

ICML 2026poster

Large language models are increasingly deployed in multi-turn interactive settings where users or environments can iteratively provide lightweight feedback. Unfortunately, optimizing such behavior presents a sharp dilemma in practice: online reinforcement learning is able to effectively address mult…

Cited by 0SourceScholar
2025

GUI-Actor: Coordinate-Free Visual Grounding for GUI Agents

NeurIPS 2025poster

One of the principal challenges in building VLM-powered GUI agents is visual grounding—localizing the appropriate screen region for action execution based on both the visual content and the textual plans. Most existing work formulates this as a text-based coordinate generation task. However, these a…

Cited by 0SourceScholar