Pratik Patil |
Research | Software | Teaching | Talks | Personal |
|
Brief bio: I am currently a postdoctoral researcher in the Department of Statistics at the University of California at Berkeley. I received my Ph.D. jointly in Statistics and Machine Learning from Carnegie Mellon University in Pittsburgh. Before that, I received an M.A.Sc. in Electrical and Computer Engineering along with an M.Sc. in Statistics from the University of Toronto, where I was an exchange student from the Indian Institute of Technology at Guwahati while pursuing a B.Tech. in Electronics and Communications Engineering. More details can be found in my CV and on my general evolutions page. Interests: Broadly in statistical machine learning, optimization, and information theory. Specifically I have on the following topics: exact asymptotics, random matrix theory, cross-validation and tuning, bagging and ensemble methods, sketching and randomized algorithms, free probability theory, uncertainty quantification, constrained statistical inference, model evaluation and benchmarking, network information theory, signal processing and wireless communications.
Email: pratikpatil[at]berkeley[dot]edu |
Asymptotically Free Ridge Ensembles: Risks, Cross-Validation, and Tuning
Pratik Patil and Daniel LeJeune
[arXiv]
Generalized Equivalences between Subsampling and Ridge Regularization
Pratik Patil and Jin-Hong Du
Neural Information Processing Systems (NeurIPS), 2023
[paper]
Mitigating Multiple Descents: A Model-Agnostic Framework for Risk Monotonization
Pratik Patil, Arun Kuchibhotla, Yuting Wei, and Alessandro Rinaldo
[arXiv]
Estimating Functionals of the Out-of-Sample Error Distribution in High-Dimensional Ridge Regression
Pratik Patil, Alessandro Rinaldo, Ryan Tibshirani
International Conference on Artificial Intelligence and Statistics (AISTATS), 2022
[paper]
Objective Frequentist Uncertainty Quantification for Atmospheric CO2 Retrievals
Pratik Patil, Mikael Kuusela, and Jon Hobbs
SIAM-ASA Journal on Uncertainty Quantification (JUQ), 2022
[paper]