Muxing Wang

PhD Candidate · Northeastern University

I am a PhD candidate in Computer Engineering at Northeastern University, advised by Prof. Lili Su. My research spans reinforcement learning, federated learning, and representation learning. Much of my work develops theoretical guarantees for learning collaboratively across heterogeneous environments and data sources, including federated Q-learning, federated actor-critic, multi-agent temporal-difference learning, and source screening for learning shared feature extractors. I have also co-authored a survey on post-training for end-to-end autonomous driving.

Before Northeastern, I received an MSc in Statistics and Operational Research from the University of Edinburgh (with Distinction), working with Prof. Daniel Paulin, and a Bachelor of Computer Science from the University of Waterloo.

News

Oct 2026 New on the blog: an interactive linear algebra visualizer that expands matrix products entry by entry (read · open).
Oct 2026 New on the blog: a horse race card game you can play against my Q-learning agent, or against a friend (play).
Sep 2026 Our survey on post-training for end-to-end autonomous driving received the Best Paper Award at the DriveX Workshop, ECCV 2026 (news).
May 2026 New preprint on collaborative yet personalized policy training with single-timescale federated actor-critic (arXiv).
Apr 2026 Our paper on source screening for learning shared feature extractors was accepted to ICML 2026 as a Spotlight.
Mar 2026 New preprint on personalized multi-agent average-reward temporal-difference learning (arXiv).
Aug 2025 Our paper on federated Q-learning across heterogeneous environments was published in TMLR.

Research Interests

Reinforcement Learning
Federated Learning
Representation Learning
Post-Training for Autonomous Driving

Publications

Accepted

Post-Training in End-to-End Autonomous Driving: Taxonomy, Methods, and Challenges

Ruining Yang^*, Muxing Wang^*, Yixiao Chen, Tongfei Guo, Yi Xu, Can Cui, Zichong Yang, Yitian Zhang, Ziran Wang, Yun Fu, Lili Su

DriveX Workshop: Foundation Models for Autonomous Driving, ECCV 2026 Oral Best Paper Award

^* Equal contribution

OpenReview · arXiv · GitHub · News
On the Power of Source Screening for Learning Shared Feature Extractors

Muxing Wang, Connor McLaughlin, Lili Su

International Conference on Machine Learning (ICML), 2026 Spotlight

arXiv
On the Convergence Rates of Federated Q-Learning across Heterogeneous Environments

Muxing Wang, Pengkun Yang, Lili Su

Transactions on Machine Learning Research (TMLR), 2025

Workshop version: International Workshop on Federated Foundation Models (FL@FM), NeurIPS 2024

OpenReview · arXiv

Under Review

Collaborative Yet Personalized Policy Training: Single-Timescale Federated Actor-Critic

Muxing Wang, Pengkun Yang, Lili Su

arXiv
Personalized Multi-Agent Average Reward TD-Learning via Joint Linear Approximation

Muxing Wang, Pengkun Yang, Lili Su

arXiv

Education

2023 – Present

Northeastern University, Boston, MA, USA

PhD Candidate in Computer Engineering · Advisor: Prof. Lili Su
2021 – 2022

University of Edinburgh, Edinburgh, UK

MSc in Statistics and Operational Research, with Distinction · Advisor: Prof. Daniel Paulin
2015 – 2020

University of Waterloo, Waterloo, Canada

Bachelor of Computer Science

Academic Services

Reviewer: ICML ARLET Workshop 2024; ICML 2026; NeurIPS 2026; ICLR 2027.