Event Calendar
Old dog, Old tricks, New show: Gradient methods for training Kernel Machines are Very Fast – Parthe Pandit
SCDLDS Colloquium
- This event has passed.
Colloquium announcement
“Old dog, Old tricks, New show: Gradient methods for training Kernel Machines are Very Fast.”
by Parthe Pandit
Thakur Family Chair Assistant Professor
Center for Machine Intelligence and Data Science (C-MInDS)
Indian Institute of Technology, Bombay
Parthe Pandit is a Thakur Family Chair Assistant Professor at the Center for Machine Intelligence and Data Science (C-MInDS) at IIT Bombay. He was a Simons postdoctoral fellow at UC San Diego. He obtained his PhD from UCLA, and his undergraduate education from IIT Bombay. He has received the AI2050 Early Career Fellowship from Schmidt Sciences in 2024, and the Jack K Wolf Student Paper Award at ISIT 2019.
Abstract: Kernel Machines are a classical family of models in Machine Learning that overcome several limitations of Neural Networks. These models have regained popularity following some landmark results showing their equivalence to Neural Networks. Folklore suggests that training procedures for Kernel Machines may not scale well for problems with large datasets, and hence Kernel Machines can only be applied when working with problems with small datasets. We dispel this belief.
After taking a fresh look at the problem of designing training algorithms for Kernel Machines, we propose a suite of algorithms based on gradient descent in the Reproducing Kernel Hilbert Space (RKHS) associated with the kernel function. These algorithms, called EigenPro, are much faster than the SOTA, and enable training of Kernel Machines with large model sizes over large datasets. This development unlocks the potential of Kernel Machines for modern applications of AI.
Based on work with Amirhesam Abedsoltan, Siyuan Ma, Yiming Zhang, and Mikhail Belkin.
Date: Tuesday, Sept 23, 2025
Time: 1:30 PM – 2:30 PM
Venue: Ramachandra Hall, AC-05 Lab-004
For details: ashoka-cdlds@ashoka.edu.in
or call: +91-9136857558
Registration and Zoom Link: /event/scdlds-coll04/
Website:
Zoom Link: https://zoom.us/j/96645811296?pwd=Q4M3qrSlVfnqjJl3flqaZa5Vvxy4Sa.1
Time: 1:30 PM – 2:30 PM
Venue: Ramachandra Hall, AC-05 Lab-004
For details: ashoka-cdlds@ashoka.edu.in
or call: +91-9136857558
Registration and Zoom Link: /event/scdlds-coll04/
Website:
Zoom Link: https://zoom.us/j/96645811296?pwd=Q4M3qrSlVfnqjJl3flqaZa5Vvxy4Sa.1