Qimao Chen (陈其茂)

Qimao Chen (陈其茂)

MLLM / Decision Intelligence

About me.

I study how multimodal decision systems can learn reliably and improve continuously in the real world.

Hi, I am a second-year master's student in the Department of Automation at Tsinghua University. My work follows a closed learning loop: constructing long-tail data, learning grounded reward signals, and post-training agents through failure feedback and interaction.

I am particularly interested in autonomous driving and embodied intelligence. I also explore quantitative research as an independent application of the same discipline: forming hypotheses from data and making decisions under uncertainty.

I believe complex systems become understandable when their patterns are observed rigorously. Good data, explicit feedback, and careful statistical analysis turn those observations into better decisions.

Feel free to reach out for discussion or collaboration.

Selected work

RESEARCH THREADS

From data and feedback to better decisions.

Across my work, the same question recurs: how can a system encounter failures, receive grounded feedback, and become more reliable after each iteration?

  1. 01

    Long-tail learning

    Constructing challenging data and scenarios for robust decision-making.

  2. 02

    Multimodal evaluation

    Learning richer reward signals from visual context, rules, and intent.

  3. 03

    Post-training agents

    Making VLA systems improve through failure feedback and interaction.

SELECTED RESEARCH

Representative work, ordered with first-author and co-first-author contributions first.

Publications and Preprints

DriveReward main figure
Preprint

DriveReward: A Comprehensive Dataset and Generative Vision-Language Reward Model for Autonomous Driving

Qimao Chen*, Fang Li*, Yuechen Luo, Zehan Zhang, Haiyang Sun, Fangzhen Li, Bing Wang, Yang Ji, Jiong Deng, Hongwei Xie, Hangjun Ye, Long Chen, Yi Zhang

arXiv
VILTA main figure
AAAI 2026

VILTA: A VLM-in-the-Loop Adversary for Enhancing Driving Policy Robustness

Qimao Chen*, Fang Li*, Shaoqing Xu*, Zhiyi Lai, Zixun Xie, Yuechen Luo, Shengyin Jiang, Hanbing Li, Long Chen, Bing Wang, Yi Zhang, Zhi-Xin Yang

arXiv
ELF-VLA main figure
CVPR 2026

Unleashing VLA Potentials in Autonomous Driving via Explicit Learning from Failures

Yuechen Luo*, Qimao Chen*, Fang Li*, Shaoqing Xu, Jiaxin Liu, Ziying Song, Zhi-Xin Yang, Fuxi Wen

arXiv
xTED main figure
AAMAS 2026

xTED: Cross-Domain Adaptation via Diffusion-Based Trajectory Editing

Haoyi Niu*, Qimao Chen*, Tenglong Liu, Jianxiong Li, Guyue Zhou, Yi Zhang, Jianming Hu, Xianyuan Zhan

arXiv GitHub
Stackelberg autonomous-background vehicle modeling main figure
NeurIPS 2023 ML4AD Workshop

Stackelberg Autonomous-Background Vehicle Modeling for Continual Policy Improvement

Haoyi Niu*, Qimao Chen*, Yingyue Li and Jianming Hu

arXiv
AdaThinkDrive main figure
ICRA 2026

AdaThinkDrive: Adaptive Thinking via Reinforcement Learning for Autonomous Driving

Yuechen Luo, Fang Li, Shaoqing Xu, Zhiyi Lai, Lei Yang, Qimao Chen, Ziang Luo, Zixun Xie, Shengyin Jiang, Jiaxin Liu, Long Chen, Bing Wang, Zhi-xin Yang

arXiv
ECCV 2026

World-in-Loop: Online Correction via Event-Triggered World Models for Robust VLA Policies

Shaoqing Xu, Fang Li, Yuechen Luo, Qimao Chen, Zhixiang Duan, Yifan Yang, Long Chen, Zhi-Xin Yang

Educations

Master's Student

Department of Automation, Tsinghua University, Beijing, China.

Undergraduate

Department of Automation, Tsinghua University, Beijing, China.

Internships

ByteDance

Post-training multimodal models for e-commerce content understanding, China.

Xiaomi EV

End-to-end autonomous driving research internship, China.