About

I am an assistant professor at the University of Iowa - Computer Science department. Before this, I was a data scientist at Foursquare. I did my PhD in Computer Science from UC Riverside, where I was fortuante to be advised by Silas Richelson and Amey Bhangale. My primary research interest lies in theoretical computer science, specifically in exploring pseudorandomness and its connections and applications to areas such as coding theory, cryptography, etc. I also like to solve problems in computer vision and worked on several applied ML/computer vision projects during my time in graduate school.

Note: I'm currently seeking PhD students to collaborate on projects within theoretical computer science. If you're interested, please send me an email. For those already at the University of Iowa wishing to discuss further, you're welcome to drop by my office.

Selected Publications

Derandomized Non-Abelian Homomorphism Testing in Low Soundness Regime, In Submission, 2024

Tushant Mittal, Sourya Roy

Analyzing Ta-Shma’s Code via the Expander Mixing Lemma, IEEE Transactions on Information Theory, 2023

Silas Richelson, Sourya Roy

Gilbert and Varshamov Meet Johnson: List-Decoding Explicit Nearly-Optimal Binary Codes, FOCS 2023

Silas Richelson, Sourya Roy

Almost Ramanujan Expanders from Arbitrary Expanders via Operator Amplification, FOCS 2022

Fernando Granha Jeronimo, Tushant Mittal, Sourya Roy, Avi Wigderson

Mixing of 3-term progressions in Quasirandom Groups, ITCS 2022

Amey Bhangale, Prahladh Harsha, Sourya Roy

Learning Long-Term Spatial-Temporal Graphs for Active Speaker Detection, ECCV 2022

Kyle Min*, Sourya Roy*,Subarna Tripathi, Tanaya Guha, Somdeb Majumdar

Locally Testable Non-Malleable Code, 2019

Silas Richelson, Sourya Roy

W-TALC: Weakly-supervised Temporal Activity Localization and Classification, ECCV '18

Sujoy Paul, Sourya Roy, Amit Roy-Chowdhury

Exploiting Transitivity for Learning Person Re-identification Models on a Budget, CVPR '18

Sourya Roy, Sujoy Paul, Neal E. Young, Amit Roy-Chowdhury

W-TALC: Weakly-supervised Temporal Activity Localization and Classification, ECCV '18

Sujoy Paul, Sourya Roy, Amit Roy-Chowdhury

Incorporating scalability in unsupervised spatio-temporal feature learning, ICASSP '18

Sujoy Paul, Sourya Roy, Amit Roy-Chowdhury