PeopleofSuperintelligence

Researcher

Fan Chen

Fan Chen is a machine-learning theorist whose research spans reinforcement learning, private regression, language-model training and sampling. He earned a bachelor’s degree at Peking University in June 2023 and coauthored research on partially observable and latent decision processes. His later papers examined private linear regression and how pre-training coverage shapes post-training; his 2026 diffusion-sampling paper won the ICML Outstanding Paper Award. As of October 2026, he is an MIT PhD student advised by Alexander Rakhlin and Constantinos Daskalakis, and he presented his sampling framework at the Simons Institute in August 2026.

Works
114
Citations
1,047
h-index
18

Publication numbers from OpenAlex, as of 22 October 2026.

Timeline

  1. June 2023

    A degree at Peking University

    In June 2023, he earns a bachelor’s degree at Peking University and works with Song Mei, Yu Bai, Zaiwen Wen and Junyu Zhang.

  2. 2023

    Limits in revealing POMDPs

    With Huan Wang, Caiming Xiong, Song Mei and Yu Bai, he coauthors the 2023 paper Lower Bounds for Learning in Revealing POMDPs. It establishes strong PAC and regret lower bounds for learning in partially observable Markov decision processes.

  3. 2024

    Learning in latent MDPs

    With Constantinos Daskalakis, Noah Golowich and Alexander Rakhlin, he coauthors Near-Optimal Learning and Planning in Separated Latent MDPs in 2024. The paper studies computational and statistical aspects of learning latent Markov decision processes and establishes a nearly sharp statistical threshold for the horizon length needed for efficient learning.

  4. 2025

    The coverage principle

    In 2025, he coauthors The Coverage Principle: How Pre-Training Enables Post-Training with Audrey Huang, Noah Golowich, Sadhika Malladi, Adam Block, Jordan T. Ash, Akshay Krishnamurthy and Dylan J. Foster. It defines coverage as the probability mass a pre-trained model places on high-quality responses and shows that coverage generalizes faster than cross-entropy.

  5. 2025

    Private linear regression

    For private linear regression, he coauthors Beyond Covariance Matrix: The Statistical Complexity of Private Linear Regression with Jiachun Li, Alexander Rakhlin and David Simchi-Levi in 2025. The paper establishes minimax convergence rates for central and local privacy and introduces Information-Weighted Regression.

  6. 2026

    Diffusion sampling wins at ICML

    In 2026, he coauthors High-accuracy sampling for diffusion models and log-concave distributions with Sinho Chewi, Constantinos Daskalakis and Alexander Rakhlin. The paper gives diffusion-sampling algorithms with an exponential improvement over previous results and the first polylogarithmic sampler for general log-concave distributions using gradient evaluations. It receives the ICML 2026 Outstanding Paper Award.

  7. August 2026

    Presenting at Simons

    At the Simons Institute for the Theory of Computing, he presents his sampling framework in August 2026.

  8. As of October 2026

    PhD student at MIT

    As of October 2026, he is a PhD student at MIT working on statistics and machine learning theory, advised by Alexander Rakhlin and Constantinos Daskalakis.

Career

  1. 2025–2026Chongqing University of Posts and Telecommunications
  2. 2019–2025Shanghai University
  3. 2010–2015Japan Advanced Institute of Science and Technology
  4. 2008–2024Southwest Jiaotong University

Education

  1. 2023Peking UniversityBachelor's degree
  2. Massachusetts Institute of TechnologyPhD

From the biography and OpenAlex affiliations (the years of papers that list the organization), so some years are approximate.

  1. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other universityIncludes Fan Chen. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other university. We compare doctorate and 5+ year employment links, career timing and current workplaces for 187 major-prize and 1,111 best-paper winners.
  2. The company share of AI researchers’ top papers was 31% in 2020–2024 and 8% in 1970–1974Includes Chongqing University of Posts and Telecommunications. We estimate where the people shaping AI worked when their most-cited papers appeared: the company share was 31% in 2020–2024 and 8% in 1970–1974, and Google led all organizations in 2020–2024.
  3. More people shaping AI studied at Massachusetts Institute of Technology than at any other institutionIncludes Massachusetts Institute of Technology. We estimate where the people shaping AI studied and where their careers led. Massachusetts Institute of Technology leads with 467 former students, including 291 recorded doctorates.
  4. Google had the most people shaping AI every year from 2014 to 2025Includes Massachusetts Institute of Technology. Our estimates of which organizations employed the most people shaping AI each year from 1955 to 2025, tracing the shift from university labs to companies.

Colleagues

People named in this biography, and people with dated positions at the same organization in the same years.

Sources and credits

Photo: We have no published photo of Fan Chen yet. Send us a free-licensed photo.

  1. Fan Chen’s homepage sites.google.com. Checked 22 October 2026.
  2. Lower Bounds for Learning in Revealing POMDPs proceedings.mlr.press. Checked 22 October 2026.
  3. Near-Optimal Learning and Planning in Separated Latent MDPs proceedings.mlr.press. Checked 22 October 2026.
  4. The Coverage Principle: How Pre-Training Enables Post-Training arxiv.org. Checked 22 October 2026.
  5. The Coverage Principle: How Pre-Training Enables Post-Training proceedings.iclr.cc. Checked 22 October 2026.
  6. Beyond Covariance Matrix: The Statistical Complexity of Private Linear Regression arxiv.org. Checked 22 October 2026.
  7. High-accuracy sampling for diffusion models and log-concave distributions arxiv.org. Checked 22 October 2026.
  8. High-Accuracy Sampling for Diffusion Models and Log-Concave Distributions proceedings.mlr.press. Checked 22 October 2026.
  9. High-Accuracy Sampling for Diffusion Models and Log-Concave Distributions simons.berkeley.edu. Checked 22 October 2026.