Matt Jordan

I am a final-year PhD student in the department of Computer Science at the University of Texas at Austin, advised by Alex Dimakis. These days I mostly work on building better datasets for pretraining of large language models. Throughout most of my PhD, I have been interested in provable verification of robustness of neural networks and other assorted challenges related to security and privacy in AI/ML. I've also dabbled in contrastive learning, generative modeling, and causality. Prior to graduate school, I received a bachelor's degree in math and computer science from MIT and cofounded a tech startup.

I am currently on the job market and am actively seeking employment!

Publications/Projects

	DataComp-LM: In search of the next generation of training sets for language models Li et. al Preprint [PDF]   [Code]
	OpenLM: a minimal but performative language modeling (lm) repository, 2023 Gururangan et. al Open Source Project   [Code]
	Conditional Generative Models are Sufficient to Sample from Any Causal Effect Estimand Md Musfiqur Rahman*, Matt Jordan*, Murat Kocaoglu ICML 2023 Workshop on Structured Probabilistic Inference & Generative Modeling. [PDF]   [Code]
	Lovasz Theta Contrastive Learning Georgios Smyrnis, Matt Jordan, Ananya Uppal, Giannis Daras, Alexandros G. Dimakis NeurIPS 2022 Workshop: Self-Supervised Learning - Theory and Practice [PDF]   [Code]
	Zonotope Domains for Lagrangian Neural Network Verification Matt Jordan*, Jonathan Hayase*, Alexandros G. Dimakis, Sewoong Oh Advances in Neural Information Processing Systems (NeurIPS) 2022. [PDF]   [Code]
	Inverse Problems Leveraging Pre-trained Contrastive Representations Sriram Ravula, Georgios Smyrnis, Matt Jordan, Alexandros G. Dimakis Advances in Neural Information Processing Systems (NeurIPS) 2021. [PDF]   [Code]
	Provable Lipschitz Certification for Generative Models Matt Jordan, Alexandros G. Dimakis International Conference on Machine Learning (ICML) 2021. [PDF]   [Code]
	Quarantines as a Targeted Immunization Strategy Jessica Hoffmann*, Matt Jordan*, Constantine Caramanis. Preprint, arXiv:2008.08262. [PDF]
	Exactly Computing the Local Lipschitz Constant of ReLU Networks Matt Jordan, Alexandros G. Dimakis. Advances in Neural Information Processing Systems (NeurIPS) 2020. [PDF]   [Code]
	Provable Certificates for Adversarial Examples: Fitting a Ball in a Union of Polytopes Matt Jordan, Justin Lewis, Alexandros G. Dimakis. Advances in Neural Information Processing Systems (NeurIPS) 2019. [PDF]   [Code]   [Talk]
	Quantifying Perceptual Distortion of Adversarial Examples Matt Jordan, Naren Manoj, Surbhi Goel, Alexandros G. Dimakis. Preprint, arXiv:1902.08265. [PDF]   [Code]