Hi! I'm Matthew Lau, a software/AI developer and a former PhD CS student from Georgia Tech.

Contact me here!

About

I love Math, and my research interests lie in data-driven methods (Machine Learning, Statistics, Causality), especially aimed at safety, trustworthiness/robustness and security. I am particularly interested answering the question: how can we know and account for the unknown? Some of my research interests along this line include anomaly/novelty detection and (adversarial) robustness. My other interests that inspire these applications include learning theory, causal modeling and implicit bias in optimization.

My theoretical work explores the intersection between anomaly/novelty detection, traditional classification and learning theory.
My applied work focuses on anomaly detection and localization through causal modeling, especially for cyber-physical security, as well as the adversarial robustness of vision-based neural networks.

Please reach out if you are interested in collaborating or chatting!

Publications

Updated list can be found on my Google scholar page.

Matthew Lau*, Tian-Yi Zhou*, Xiangchi Yuan, Jizhou Chen, Wenke Lee, Xiaoming Huo. Bridging Unsupervised and Semi-Supervised Anomaly Detection: A Theoretically-Grounded and Practical Framework with Synthetic Anomalies. In submission.
Matthew Lau*, Haoran Wang*, Alec Helbling, Matthew Hull, ShengYun Peng, Mansi Phute, Martin Andreoni, Willian T. Lunardi, Duen Horng Chau, Wenke Lee. ARMOR: Manifold-Oriented Training for Adversarially Robust Aerial Object Detection under Data Scarcity. In submission.
Matthew Daniel Hull, Haoyang Yang, Pratham Mehta, Mansi Phute, Aeree Cho, Haoran Wang, Matthew Lau, Wenke Lee, Willian T. Lunardi, Martin Andreoni, Duen Horng Chau. 3D Gaussian Splat Vulnerabilities. CVPR Workshop 2025 (Neural Fields Beyond Conventional Cameras).
Matthew Hull, Haoran Wang, Matthew Lau, Alec Helbling, Mansi Phute, Chao Zhang, Zsolt Kira, Willian Lunardi, Martin Andreoni, Wenke Lee, and Polo Chau. RenderBender: A Survey on Adversarial Attacks Using Differentiable Rendering. IJCAI 2025.
Tian-Yi Zhou*, Matthew Lau*, Jizhou Chen, Wenke Lee, Xiaoming Huo. Optimal Classification-based Anomaly Detection with Neural Networks: Theory and Practice. In submission.
Matthew Lau, Fahad Alsaeed, Kayla Thames, Nano Suresettakul, Saman Zonouz, Wenke Lee, Athanasios P Meliopoulos. Physics-Assisted Explainable Anomaly Detection in Power Systems. ECAI 2024.
Matthew Lau, Haoran Wang, Alec Helbling, Matthew Hull, ShengYun Peng, Willian T. Lunardi, Martin Andreoni, Wenke Lee. Non-Robust Features are Not Always Useful in One-Class Classification. CVPR Workshop 2024 (Visual Anomaly and Novelty Detection Workshop).
Matthew Lau, Leyan Pan, Stefan Davidov, Athanasios P Meliopoulos, Wenke Lee. Geometric Implications of Classification on Reducing Open Space Risk. ICLR Tiny Paper 2024 (Invited to Present).
Matthew Lau, Ismaila Seck, Athanasios P Meliopoulos, Wenke Lee, Eugene Ndiaye. Revisiting Non-separable Binary Classification and its Applications in Anomaly Detection. TMLR 2024.
Matthew Lau, Kayla Thames, Athanasios P Meliopoulos. Active Distribution System Coordinated Control Method via Artificial Intelligence. 11th Bulk Power Systems Dynamics and Control Symposium (IREP) 2022.

Legend: * denotes equal contribution.

Other Projects

Apart from these publications, I have also collaborated on and been involved in projects on:

Physics-informed Neural Networks (power flow and PDE based)
Large Language Model (LLM) Reasoning: Theory, Application, Dataset
LLM Agent Security
Drone Swarm Anomaly Detection

For Fun

In my former life, I used to play chess competitively! I've played in tournaments in about 8 countries and still play for fun whenever I have the chance :)

Page updated

Google Sites

Report abuse