Lei Song

Principal Researcher, Microsoft Research Asia, Microsoft

Joined

May 2021

Names

Lei Song (Preferred)

Emails

****@microsoft.com (Confirmed)

****@gmail.com (Confirmed)

Personal Links

Google Scholar

DBLP

Semantic Scholar

Career & Education History

Principal Researcher

Microsoft Research Asia, Microsoft (microsoft.com)

2020 – Present

Advisors, Relations & Conflicts

Coworker

Jiang Bian

2020 – 2020

PhD Advisor

Holger Hermanns

2013 – 2015

Expertise

Reinforcement learning

2018 – Present

Model checking

2014 – Present

Publications

In-Context Compositional Q-Learning for Offline Reinforcement Learning
Qiushui Xu, Yu-Hao Huang, Yushu Jiang, Wenliang Zheng, Lei Song, Jinyu Wang, Jiang Bian
- ICLR 2026 Poster
- Readers: Everyone
Holdout-Loss-Based Data Selection for LLM Finetuning via In-Context Learning
Ling Zhang, Xianliang Yang, Juwon Yu, Park Cheonyoung, Lei Song, Jiang Bian
- ICLR 2026 Poster
- Readers: Everyone
Sample-efficient LLM Optimization with Reset Replay
Zichuan Liu, Jinyu Wang, Lei Song, Jiang Bian
- Submitted to ICLR 2026
- Readers: Everyone
PixelCraft: A Multi-Agent system for High-Fidelity Visual Reasoning on Structured Images
Shuoshuo Zhang, Zijian Li, Yizhen Zhang, Jingjing Fu, Lei Song, Jiang Bian, Jun Zhang, Yujiu Yang, Rui Wang
- ICLR 2026 Poster
- Readers: Everyone
Rethinking Reward Models for Multi-Domain Test-Time Scaling
Dong Bok Lee, Seanie Lee, Sangwoo Park, Minki Kang, Jinheon Baek, Dongki Kim, Dominik Wagner, Jiongdao Jin, Heejun Lee, Tobias Bocklet, Jinyu Wang, Jingjing Fu, Sung Ju Hwang, Jiang Bian, Lei Song
- Submitted to ICLR 2026
- Readers: Everyone
NaDRO: Leveraging Dual-Reward Strategies for LLMs Training on Noisy Data
Haolong Qian, Xianliang Yang, Ling Zhang, Lei Song, Jiang Bian, Chun Yuan
- NeurIPS 2025 poster
- Readers: Everyone
Knowing What Not to Do: Leverage Language Model Insights for Action Space Pruning in Multi-agent Reinforcement Learning
Zhihao Liu, Xianliang Yang, Zichuan Liu, Yifan Xia, Wei Jiang, Yuanyu Zhang, Lijuan Li, Guoliang Fan, Lei Song, Jiang Bian
- Accepted by TMLR
- Readers: Everyone
Unveiling Markov heads in Pretrained Language Models for Offline Reinforcement Learning
Wenhao Zhao, Qiushui Xu, Linjie Xu, Lei Song, Jinyu Wang, Chunlai Zhou, Jiang Bian
- ICML 2025 poster
- Readers: Everyone
From Complex to Atomic: Enhancing Augmented Generation via Knowledge-Aware Dual Rewriting and Reasoning
Jinyu Wang, Jingjing Fu, Rui Wang, Lei Song, Jiang Bian
- ICML 2025 poster
- Readers: Everyone
HeurAgenix: A Multi-Agent LLM-Based Paradigm for Adaptive Heuristic Evolution and Selection in Combinatorial Optimization
Xianliang Yang, Lei Song, Yapu Zhang, Jiang Bian
- ICLR 2025 Conference Withdrawn Submission
- Readers: Everyone

View all 24 publications

Co-Authors

View all 73 co-authors