Embedded Files
Hello!
Hello!
I am a PhD student in the Department of Electrical and Computer Engineering at the University of California, Santa Barbara. My current research interests are principled design of deep learning, imbalanced learning in overparameterized models. I am fortunate to have Prof Christos Thrampoulidis as my PhD advisor. Please see here for a summary of my selected research work.
Prior to joining UCSB in Fall 2019, I worked as a Senior Communication System Engineer at Maxlinear in Bangalore, where I was responsible for design, mathematical analysis and practical implementation of adaptive signal processing algorithms for 5G transceiver.
I have a Masters in ECE from the Indian Institute of Science (IISc), during which my research was on Information Theory for Distributed Storage under the guidance of Prof P. Vijay Kumar.
Connect with me on LinkedIn here!
News
News
[Nov 2023] Contributed talk at the 2023 Conference on the Mathematical Theory of Deep Neural Networks (DeepMath)
[Sept 2023] Completed an exciting internship as an Applied Scientist II Intern at Amazon Lab126! Developed Deep Learning and Deep Reinforcement Learning models to solve signal processing problems in the speech/audio space.
[Dec 2022] NeurIPS 2022: presented "Imbalance Trouble: Revisiting Neural Collapse Geometry" at the conference and a follow up work at the OPT2022 workshop at NeurIPS. Thank you all for visiting our posters and the insightful conversations!
[Oct 2022] Invited talk at Prof Misha Belkin's group (UC San Diego) on "imbalance trouble in overparameterized learning"
[Oct 2022] Our recent work on implicit geometry of CE parameterizations accepted at OPT workshop, NeurIPS 2022!
[Sept 2022] Our paper "Imbalance Trouble: Revisiting Neural Collapse Geometry" accepted at NeurIPS 2022!
[Dec 2021] NeurIPS 2021: presented our paper on "Label Imbalanced and Group-Sensitive Classification under Overparameterization"
[Oct 2021] Invited talk at an IEEE event at IISc
[April 2021] Gave a talk at the Workshop on the Theory of Overparameterized Machine Learning - TOPML2021 on our recent work on Label Imbalanced and Group-Sensitive Classification under Overparameterization. Video recording will be available on the workshop website soon!
[Jan 2021] Our paper on high-dimensional multiclass linear classification has been accepted for presentation at the 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) and publication in the proceedings.
[March 2020] Our paper "Analytic Study of Double Descent in Binary Classification: The Impact of Loss " has been accepted for presentation at the 2020 IEEE International Symposium on Information Theory (ISIT) and for publication in the proceedings.
[March 2020] Our proposal has been selected as a Finalist for the Qualcomm Innovation Fellowship 2020 North America.
Page updated
Google Sites
Report abuse