Parthe Pandit

Mood Indigo 2015 

HDSI-Simons Postdoctoral Fellow
Halıcıoğlu Data Science Institute
UC San Diego

Email: <firstname> <lastname> [at] ucsd [dot] edu

About Me

I am a postdoc at UCSD where I work with Misha Belkin and Arya Mazumdar. My research focusses on understanding the interplay between generalization and optimization in Machine Learning.

I defended my Ph.D. thesis in ECE at UCLA in 2021 where I worked with Alyson Fletcher, Sundeep Rangan, and Arash Amini. I also have an MS in Statistics from UCLA, and a B.Tech.+M.Tech. degree in EE from IIT Bombay, with a minor in CS. Between 2015-2017 I worked with Sam Coogan, Ankur Kulkarni, and Preeti Rao.

News

  • [June 2022] Libin Zhu, Misha Belkin and I wrote a note on the bilinear nature of bottleneck networks arXiv 2206.15058

  • [May 2022] Daniel Beaglehole, Misha Belkin and I submitted a paper on inconsistency of kernel interpolation in low dimensions arXiv 2205.13525

  • [Jan 2022] New paper on hidden linearity of kernels and multi-Layer perceptrons: arXiv 2201.08082

  • [Jan 2022] I am now a Postdoc at the Halıcıoğlu Data Science Institute at UC San Diego.

  • [Dec 2021] Submitted MS thesis titled “Non-asymptotic Analysis of Learning Long-range Autoregressive Generalized Linear Models for Discrete High-dimensional Data”.

  • [Nov 2021] I defended my PhD dissertation titled “Exact analysis of Inverse problems in High dimensions with Applications to Machine Learning”.

  • [June 2021] I am a Quantitative Researcher Intern at Citadel LLC. working with Hua Zheng, Julius Bonart, and Paul Jefferys from the CLB Trading team.

  • [May 2021] Paper accepted at ICML 2021: Implicit Bias of Linear RNNs.

  • [Mar 2021] I have been awarded the HDSI-Simons Postdoctoral Fellowship

  • [Nov 2020] Our journal paper Generalized Autoregressive Linear Models for Discrete High-dimensional Data was accepted at the IEEE Journal on Special topics in Information Theory. This was a special issue on Estimation and Inference.