PeopleofSuperintelligence

Researcher

Muhan Zhang

Also known as 张牧涵

Muhan Zhang is a computer scientist at Peking University whose work on graph neural networks produced SortPooling for graph classification and SEAL for link prediction. He completed a computer science PhD at Washington University in 2019 under Yixin Chen, then worked as a research scientist at Facebook AI from 2019 to 2021. In 2021, he joined the Institute for Artificial Intelligence at Peking University, and in 2024 he and his coauthors proposed PiSSA, a fine-tuning method for large language models, and GOFA, a generative graph language model. A 2025 paper he coauthored, "Reconsidering the Performance of GAE in Link Prediction", won the CIKM-25 Best Full Paper Award. He is an assistant professor and assistant to the dean at the institute, and his research covers self-evolving and lifelong AI together with graph machine learning.

Badges

Timeline

  1. 2015–2019

    PhD at Washington University

    He graduates from the IEEE pilot class of Shanghai Jiao Tong University in 2015 and earns a computer science PhD at Washington University in St. Louis in 2019, advised by Yixin Chen.

  2. 2018–2020

    SortPooling and SEAL

    In 2018, he publishes SortPooling for graph classification at AAAI, where it is selected as one of the top ten most influential papers of AAAI-2018, and, with Yixin Chen, the link prediction method SEAL at NeurIPS. In 2020, SEAL takes first place on three of four OGB leaderboard link prediction tasks.

  3. 2019–2021

    Research scientist at Facebook AI

    He works as a research scientist at Facebook AI (now Meta AI) from 2019 to 2021, on large-scale graph learning systems and problems.

  4. May 2021

    Joining Peking University

    In May 2021, he joins Peking University as an assistant professor at its Institute for Artificial Intelligence.

  5. 2022–2023

    Textbook chapter and AI 2000 honors

    The textbook Graph Neural Networks: Foundations, Frontiers, and Applications is published in 2022, and he contributes its chapter on link prediction. AMiner gives him AI 2000 most influential scholar honorable mentions in 2022 and 2023.

  6. 2024

    PiSSA and GOFA

    In April 2024, he and coauthors Fanxu Meng and Zhaohui Wang propose PiSSA, a parameter-efficient fine-tuning method that initializes LoRA-style adapters with the principal components of a model's weight matrix. In July 2024, he and Yixin Chen, with other coauthors, propose GOFA, a generative graph language model.

  7. 2025

    CIKM best full paper award

    In 2025, he coauthors the paper "Reconsidering the Performance of GAE in Link Prediction" with three collaborators, and it wins the CIKM-25 Best Full Paper Award.

  8. As of October 2026

    Assistant professor at Peking University

    He is an assistant professor and assistant to the dean at Peking University's Institute for Artificial Intelligence. His research covers self-evolving and lifelong AI, including In-Parameter Learning that turns transient context into long-term memory, together with graph machine learning, and six papers from his group are accepted at NeurIPS-26 in September 2026.

Career

  1. Since 2021Peking UniversityAssistant Professor
  2. 2019–2021MetaResearch Scientist

Education

  1. 2019Washington University in St. LouisPhD
  2. 2015Shanghai Jiao Tong University

From the biography.

  1. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other universityIncludes Muhan Zhang. 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. 23% of people shaping AI who left a university for a company later returned to a universityIncludes Muhan Zhang. We estimate that 468 of the 2,011 people shaping AI who left a university for a company (23%) later returned to a university after a median of 5 years, and 163 of 829 (20%) made the reverse round trip.
  3. The median h-index is 48 at universities and 14 at frontier AI labsIncludes Peking University. The median h-index is 48 at universities and 14 at frontier AI labs. We compare publishing careers, recent citations, dual appointments and job moves across 2,420 university researchers and 308 frontier-lab researchers.
  4. The company share of AI researchers’ top papers was 31% in 2020–2024 and 8% in 1970–1974Includes Peking University. 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.

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 Muhan Zhang yet. Send us a free-licensed photo.

  1. Muhan Zhang - Institute for Artificial Intelligence, Peking University ai.pku.edu.cn. Checked 21 October 2026.
  2. Muhan Zhang's Homepage muhanzhang.github.io. Checked 21 October 2026.
  3. An end-to-end deep learning architecture for graph classification - WashU Research Profiles profiles.wustl.edu. Checked 21 October 2026.
  4. Link Prediction Based on Graph Neural Networks - NeurIPS Proceedings papers.nips.cc. Checked 21 October 2026.
  5. Graph Neural Networks: Foundations, Frontiers, and Applications - GNNBook graph-neural-networks.github.io. Checked 21 October 2026.
  6. PiSSA: Principal Singular Values and Singular Vectors Adaptation of Large Language Models arxiv.org. Checked 21 October 2026.
  7. GOFA: A Generative One-For-All Model for Joint Graph Language Modeling arxiv.org. Checked 21 October 2026.