Yutong Xie   谢雨桐

I am a Ph.D. candidate in the School of Information at the University of Michigan, where I work with Prof. Qiaozhu Mei as a member of the Foreseer Research Group. Prior to this, I received my Bachelor's degree from Shanghai Jiao Tong University as a member of the ACM Honors Class, where I was advised by Prof. Yong Yu and Prof. Weinan Zhang.

I have a general research interest in exploring the potential of AI to understand, assist, and drive innovation, with a specific focus on both scientific innovation (AI for science) [ICLR'21, ICLR'23] and creative endeavors (AI for creativity) [WWW'23]. My research encompasses three key aspects:

  1. Identifying the innovation space ("What to innovate") [WWW'23];
  2. Devising computational methodologies for innovative solutions ("How to innovate") [ICLR'21]; and
  3. Establishing robust criteria for evaluating these innovations ("How to evaluate") [ICLR'23].
A significant part of my work involves examining the synergistic relationship between AI and humans under the context of innovation as well as broader scenarios. Particularly, I am interested in the problems of how to understand, align, and direct the behaviors of powerful AI such as large language models (AI behavioral science) [PNAS'24], which is crucial for fostering better human-AI collaborations.

  For more information, please check my curriculum vitae.
  yutxie AT umich DOT edu, yutongxie98 AT gmail DOT com
  Google Scholar, Semantic Scholar, Github, Twitter, LinkedIn

NEW! Our paper on a behavirol Turing test of ChatGPT was published at PNAS!
NEW! We made a video that shows the evolution of AI and data mining over half a century!
NEW! Our paper on text-to-image prompt analysis was accepted by WWW 2023 in the Creative Web Track!
NEW! Our paper on chemical space coverage measures was accepted by ICLR 2023!

Selected Publications

For the full publication list, please refer to my Google Scholar profile.

A Turing test of whether AI chatbots are behaviorally similar to humans
Qiaozhu Mei, Yutong Xie, Walter Yuan, Matthew O. Jackson.
Proceedings of the National Academy of Sciences of the United States of America (PNAS), Feb 2024.
[arXiv] [SSRN] [Code] [UMich News] [Stanford News]

A Prompt Log Analysis of Text-to-Image Generation Systems
Yutong Xie*, Zhaoying Pan*, Jinge Ma*, Luo Jie, Qiaozhu Mei.
The ACM Web Conference (WWW), 2023 (The Creative Web Track).
[PDF] [arXiv] [Code] [Video] [Slides]

How Much Space Has Been Explored? Measuring the Chemical Space Covered by Databases and Machine-Generated Molecules
Yutong Xie, Ziqiao Xu, Jiaqi Ma, Qiaozhu Mei.
International Conference on Learning Representations (ICLR), 2023.
International Conference on Machine Learning (ICML) AI for Science Workshop, 2022.
[PDF] [Code] [SlidesLive] [Slides] [Poster] [AI Time Talk]

Multi-View Graph Representation for Programming Language Processing: An Investigation into Algorithm Detection
Ting Long*, Yutong Xie*, Xianyu Chen, Weinan Zhang, Qinxiang Cao, Yong Yu.
AAAI Conference on Artificial Intelligence (AAAI), 2022 (acceptance rate 15%).
[PDF] [Code] [SlidesLive] [Slides] [Poster]

MARS: Markov Molecular Sampling for Multi-objective Drug Discovery
Yutong Xie, Chence Shi, Hao Zhou, Yuwei Yang, Weinan Zhang, Yong Yu, Lei Li.
International Conference on Learning Representations (ICLR), 2021.
Spotlight presentation (top 5%).
[PDF] [Code] [SlidesLive] [Slides] [Poster] [AI Time Talk] [WeChat Article]

* = equal contribution