Poster · Sydney
Beyond Outcome Rewards: Step-Level Self-Distilled Policy Optimization for Deep Search Agents
- When
- Tue 8 Dec 2026, 10:00–13:00 AEDT
- Where
- Hall 1-4, ICC Sydney
Exhibition Centre, Level 1 - Session
- Sydney Poster Session 1
- Authors
- Haoze Wu, Chuqiao Kuang, Tianyi Zhuang, Xiaoguang Li
- Organisations
- Hong Kong University of Science and Technology, Alibaba, Huawei
- QRS classification
- Reinforcement learning, Agentic AI