Wendi Yu

Wendi Yu

I am a Ph.D. student in Computer Science at Texas A&M University, advised by Prof. Shuiwang Ji. Previously, I received my bachelor’s degree in Software Engineering from Tongji University and worked as a research intern with Prof. Xiaorui Liu at North Carolina State University.

My research focuses on efficient diffusion language models and agentic reasoning. I investigate the principles that enable effective reasoning within language models and collaboration across agents, and develop methods for applying these capabilities to scientific discovery.

Google ScholarCVLinkedIn

news

  • Two papers accepted at NeurIPS 2026, including my first-author work on multi-agent systems. [Multi-agent systems] [Equivariant networks]
  • Joined as a Graduate Research Fellow at the Texas A&M Institute of Data Science (TAMIDS).
  • Our collaborative work on deep learning and amyotrophic lateral sclerosis was published in Brain. [Paper]
  • My first-author paper, “Automated Polynomial Filter Learning for Graph Neural Networks,” was accepted at IEEE BigData 2024. [Paper]
  • Started my Ph.D. in Computer Science at Texas A&M University.

Selected Publications

(* equal contribution)
AMEvo framework: individual agent summaries, memory synthesis, and asymmetric updates to shared and private memory.

Multi-Agent Asymmetric Memory Evolution for Algorithm Design

Lianhao Zhou, Wendi Yu, Brady Nguyen, Declan Staunton, Sai Sudarshan Barath, Shuiwang Ji

Manuscript, 2026

All publications →