Tianyu Zhang

Tianyu Zhang

Ph.D. Student in Machine Learning

MILA - Québec Institute for Artificial Intelligence

Biography

Tianyu Zhang is a Ph.D. student supervised by Prof. Yoshua Bengio at Mila. He leads the RICE-N project which wins Netexplo Innovation Award in 2023. His research interests include Algorithmic Game Theory, Agent-based Model Simulator, AI for Climate Change, Multi-agent Reinforcement Learning, Modular Deep Learning etc. Previously, he worked as a quantitative researcher in financial engineering. Please feel free to contact him if you are interested in his work.

Download the CV.

Education
  • Ph.D Student in Machine Learning, 09/2021 - Present

    MILA - Québec Institute for Artificial Intelligence

  • M.Sc. in Machine Learning, 09/2020 - 06/2021

    MILA - Québec Institute for Artificial Intelligence

  • M.S. in Applied Mathematics, 09/2018 - 01/2020

    Courant Institute at New York University

  • B.S. in Mathematics and B.Econ. in Finance, 09/2014 - 06/2018

    Wuhan University

Recent Publications

Quickly discover relevant content by filtering publications.
(2025). STRICT: Stress Test of Rendering Images Containing Text. In arXiv 2505.18985.

Cite

(2025). Advantage Alignment Algorithms. In ICLR.

Cite

(2025). MuPT: A Generative Symbolic Music Pretrained Transformer. In ICLR 2025.

Cite

(2025). MAP: Low-compute Model Merging with Amortized Pareto Fronts via Quadratic Approximation. In ICLR 2025.

Cite

(2025). VCR: A Task for Pixel-Level Complex Reasoning in Vision Language Models via Restoring Occluded Text. In ICLR.

Cite

(2024). BigDocs: An Open and Permissively-Licensed Dataset for Training Multimodal Models on Document and Code Tasks. In ICLR 2025.

Cite

(2024). Expected flow networks in stochastic environments and two-player zero-sum games. In International Conference on Learning Representations 2024.

Cite

(2023). Large-scale Contrastive Language-Audio Pretraining with Feature Fusion and Keyword-to-Caption Augmentation . In ICASSP.

PDF Cite Code Dataset

(2023). Multi-Agent Reinforcement Learning for Fast-Timescale Demand Response of Residential Loads. In AAMAS 2023 Extened Abstract.

Cite

(2022). Biological Sequence Design with GFlowNets. In ICML.

Cite

(2022). (Private)-Retroactive Carbon Pricing [(P)ReCaP]: A Market-based Approach for Climate Finance and Risk Assessment. In arXiv.org.

Cite

(2021). ClimateGAN: Raising Climate Change Awareness by Generating Images of Floods. In ICLR.

Cite

(2017). A minimal sufficient set of procedures in a bargaining model. In Economics Letters.

Cite

Contact