This blog includes some thoughts on recent fantastic works in machine learning, natural language processing and applied mathematics.

Reading maketh a full man; conference a ready man; and writing an exact man.

About me

Hi, I am Yimeng Chen (陈奕梦), a fourth year Ph.D. student in Statistics at Academy of Mathematics and Systems Science, CAS. Before that, I recieved my Bachelor degree in Physics at the University of Chinese Academy of Science. Up to now, I have 1 years of research experience in Physics and 3 years in Natural Language Processing and Machine Learning.

I like to think about the general mechanics behind algorithms in different forms. I am seeking for designing machine learning algorithms that are both practical and well grounded by theories.

Contact

I am open to academic discussion and coorperation, and also looking forward to visiting and job oppotunities. Please feel free to contact me if you are interested.

Reseach interests

  • Out-of-distribution generalization.
  • Uncertainty calibration and out-of-distribution detection.

Publications

When Does Group Invariant Learning Survive Spurious Correlations?
Yimeng Chen, Ruibin Xiong, Zhiming Ma and Yanyan Lan. (NeurIPS 2022)

PEMP: Leveraging Physics Properties to Enhance Molecular Property Prediction
Yuancheng Sun*, Yimeng Chen*, Weizhi Ma, Wenhao Huang, Kang Liu, Zhiming Ma, Wei-Ying Ma and Yanyan Lan. (CIKM 2022)

Uncertainty Calibration for Ensemble-Based Debiasing Methods [Poster] [Slide]
Ruibin Xiong*, Yimeng Chen*, Liang Pang, Xueqi Cheng, Zhiming Ma and Yanyan Lan. (NeurIPS 2021)

Evaluating Natural Language Generation via Unbalanced Optimal Transport [Poster] [Slide] [Video]
Yimeng Chen, Yanyan Lan, Ruibin Xiong, Liang Pang, Zhiming Ma and Xueqi Cheng. (IJCAI 2020)

Extraction of Tight Binding Parameters From In-situ ARPES on the Continuously Doped Surface of Cuprates
Yigui Zhong*, Yimeng Chen*, Jianyu Guan, Jin Zhao, Zhicheng Rao, Cenyao Tang, H. Liu, Yujie Sun and Hong Ding. Science China Physics, Mechanics & Astronomy 61 (2018): 1-5.

Open-Source Projects

  • Lazy EMD: the Lazy Earth Mover’s Distance, which relaxes the hard constraints on the marginals in EMD. It has better matching performance when used as the distance between sentences.
  • Awesome-AI4Drug-Papers: A collection of AI for Drug Design related papers (2021-2022).
  • Awesome: anti-Shortcut Learning: A curated list of ‘anti-shortcut learning’, ‘combating spurious correlations/features/attributes’, and ‘debiasing’ related papers. In progress.

Slides & Notes

A curated list of my slides and notes on some interesting topics.

  • A paper-list of AI for Healthcare (2018-2021.03): link
  • A talk on Invariance, Causality and Robustness, in 2020: link
  • A talk on Causal Effects Idenfication with PAG, in 2020: link
  • A comprehensive review of applications of optimal transport in maching learning (2004-2017) (in a list form): link
  • A talk on WGAN, in 2017: link

Education

  • Academy of Mathematics and Systems Science, CAS. Sep 2018-Present
  • University of Chinese Academy of Science, Sep 2014-July 2018
    • Bachelor of Science in Physics
    • Supervisor: Prof. Hong Ding