Email: gouthamr [at] cmu [dot] edu
Hi, I'm Goutham Rajendran.
I am a postdoctoral researcher in the Machine Learning Department at Carnegie Mellon University, working with Pradeep Ravikumar. I recently graduated with a PhD in Computer Science from the University of Chicago, where I was extremely fortunate to have been advised by Madhur Tulsiani (Toyota Technological Institute at Chicago) and Aaron Potechin (CS @ UChicago). I've also had the pleasure of working with Bryon Aragam (UChicago Booth School of Business) and Aravindan Vijayaraghavan (Northwestern University).
Recently, I've grown deeply interested in and have conducted research on representation learning and latent variable modeling. These days, I think about causal representation learning; in particular, given raw, unstructured data, can we learn the generative model that fits the data and on top of this, also learn the causal relationships among the learnt latent variables? This is driven by the motivation to learn representations of data that are robust, explainable and fair.
My main PhD research was on understanding the optimality of the Sum-of-Squares hierarchy, one of the most powerful techniques in convex optimization, for various problems in machine learning, robust statistics, statistical physics, etc. This involves understanding the behavior of random matrices, which is a fascinating topic in its own right. I still actively work on problems in this domain.
Sum-of-Squares Lower Bounds for Densest k-Subgraph
Chris Jones, Aaron Potechin, Goutham Rajendran*, Jeff Xu
Symposium on Theory of Computing (STOC) 2023
Concentration of polynomial random matrices via Efron-Stein inequalities
Goutham Rajendran*, Madhur Tulsiani
Symposium on Discrete Algorithms (SODA) 2023 [arXiv]
Identifiability of deep generative models without auxiliary information
Bohdan Kivva*, Goutham Rajendran*, Pradeep Ravikumar, Bryon Aragam
Workshop: Causal Representation Learning at Uncertainty in Artificial Intelligence (UAI) 2022
Conference: Neural Information Processing Systems (NeurIPS) 2022 (Spotlight) [arXiv]
Sub-exponential time Sum-of-Squares lower bounds for Principal Components Analysis
Aaron Potechin, Goutham Rajendran*
Neural Information Processing Systems (NeurIPS) 2022
Nonlinear Random Matrices and Applications to the Sum of Squares Hierarchy
Goutham Rajendran
PhD Dissertation, 2022, University of Chicago [arXiv] [Slides]
Structure learning in polynomial time: Greedy algorithms, Bregman information and exponential families
Goutham Rajendran, Bohdan Kivva, Ming Gao, Bryon Aragam
Neural Information Processing Systems (NeurIPS) 2021 [arXiv] [Slides] [Poster]
Learning latent causal graphs via mixture oracles
Bohdan Kivva, Goutham Rajendran, Pradeep Ravikumar, Bryon Aragam
Neural Information Processing Systems (NeurIPS) 2021 [arXiv] [Slides (job talk)]
Sum-of-Squares Lower Bounds for Sparse Independent Set
Chris Jones, Aaron Potechin, Goutham Rajendran*, Madhur Tulsiani, Jeff Xu
Foundations of Computer Science (FOCS) 2021 [arXiv]
Sum-of-Squares Lower Bounds for Sherrington-Kirkpatrick via Planted Affine Planes
Mrinalkanti Ghosh, Fernando Granha Jeronimo, Chris Jones, Aaron Potechin, Goutham Rajendran*
Foundations of Computer Science (FOCS) 2020 [arXiv]
Combinatorial Optimization via the Sum of Squares Hierarchy
Goutham Rajendran
Master's thesis, 2018, University of Chicago [arXiv] [Slides]
Analyzing Robustness of End-to-End Neural Models for Automatic Speech Recognition
Goutham Rajendran*, Wei Zou
Manuscript [arXiv]
Machinery for Proving Sum-of-Squares Lower Bounds on Certification Problems
Aaron Potechin, Goutham Rajendran*
Manuscript [arXiv]
I used to do a lot of competitive programming, including ICPC. My handle in online judges is xorfire: Codeforces, Topcoder, Codechef.
In another life, I would have been a professional footballer (read: soccer) but in this one, my career has been plagued with injuries :)
I avidly play and watch football, especially Futbol Club Barcelona games. Recently, I have been quite interested in the NBA.
Website theme based on minimal by orderedlist