Grigorios Chrysos

Assistant Professor
University of Wisconsin-Madison
chrysos [@] wisc.edu
Madison, WI
+1 (608) 2637768
Grigorios Chrysos

About

My name is Grigoris and I am an Assistant Professor in University of Wisconsin-Madison. My research focuses on trustworthy machine learning. I am particularly passionate about developing innovative algorithms and techniques that can understand and interpret data in complex real-world scenarios.

I am recruiting motivated PhD students from the ECE or CS program. If you have already been admitted to UW-Madison ECE/CS PhD program and are interested in my research, I encourage you to reach out.

If you are in Madison and are interested to learn more on trustworthy machine learning, I am teaching ECE/Stat/Math 888 next semester - we will have a significant component dedicated to trustworthy machine learning, so feel free to check it out.

Research Interests

Machine Learning

Deep learning architectures, theoretical foundations of machine learning with applications to real-world problems.

Trustworthy Machine Learning

Robustness to malicious attacks, reasoning under uncertainty, privacy, algorithms that can adapt to changing environments.

Parsimonious Learning

Novel architectures, training methodologies, and efficient learning paradigms.

Research Highlights

MoE

Multilinear Mixture of Experts: Scalable Expert Specialization through Factorization

Can we scale the number of experts to thousands?

Paper Code Project
NeurIPS 2024
Robustness

Revisiting character-level adversarial attacks for Language Models

How effective are simple character-level attacks in modern language models?

Paper Code
ICML 2024
Diffusion models

Going beyond compositional generalization, DDPMs can produce zero-shot interpolation

Can diffusion models synthesize content they have never seen during training?

Paper Code
ICML 2024

News

Service

Service Awards

Area Chair

Area Chair in NeurIPS ('24, '25), ICLR ('25, '26), ICML ('25).
Action Editor in TMLR.

Tutorials, Workshops and Special Issues

Workshop titled "Reliable Agentic AI: From Hallucination to Trustworthy Autonomy" in ICLR'26.
Special Issue on "Post-training in Large language models for computer vision" in the prestigious International Journal of Computer Vision.
Tutorial on "Foundations of Tensor Computations for AI" in NeurIPS'25. The recording is available here.
Dagstuhl international seminar on "Tensor Factorizations Meet Probabilistic Circuits".
Workshop titled "Findings of ICCV" in ICCV'25.
Workshop titled "Another Brick in the AI Wall: Building Practical Solutions from Theoretical Foundations" in CVPR'25.
Workshop titled "Quantify Uncertainty and Hallucination in Foundation Models: The Next Frontier in Reliable AI" in ICLR'25. The recording is available here.
Workshop titled "CoLoRAI - Connecting Low-Rank Representations in AI" in AAAI'25.
Workshop titled "Fine-Tuning in Modern Machine Learning" in NeurIPS'24.
Tutorial on "Architecture Design: From Neural Networks to Foundation Models" in IEEE International Conference on Data Science and Advanced Analytics (DSAA)'24.
Tutorial on "Scaling and Reliability Foundations in Machine Learning" in IEEE International Symposium on Information Theory (ISIT)'24.
Tutorial on "Deep Learning Theory for Vision Researchers" in CVPR'23.
Tutorial on "Polynomial nets in deep learning architectures" in AAAI'23.
Tutorial on "High-degree polynomial networks for image generation and recognition" in CVPR'22.

Funding

I would like to acknowledge the funding of the following organizations who have generously supported various events or projects in the past. I am very thankful for their support:

2025
Zulip
Sponsored hosting from Zulip, which is an open-source team collaboration tool.
2024
Google and OpenAI
Grants on trustworthy Large Language Models (LLMs).
2024
ELISE Fellows Mobility Program
Travel grant for short-term visit of an ELLIS lab.