I delivered a talk in the Algebraic Geometry and Machine Learning Session, geared towards understanding distillation and exploring the possibility of structured distillation algorithms based on our NeurIPS paper “Should Under-parameterized Student Networks Copy or Average Teacher Weights?”
I attended NeurIPS in person for the first time, and presented a poster on our new paper on under-parameterized neural nets.
I gave a lecture on landscape complexity of neural networks at the Analytical Approaches for Neural Network Dynamics Workshop at Institut Henri Poincaré. Yay!
I started as a Faculty Fellow at NYU CDS and presented findings from my Ph.D. thesis at the lunch seminar.
I attended the Spin Glass Workshop at the SwissMAP Research Station in Les Diablerets and presented a poster.
Young researchers from Switzerland and Japan presented their research. Finally I had the pleasure of giving a talk in person. See the workshop website for talks including mine.
I visited Meta AI Paris for a week to meet with the team I worked with during my remote internship in the summer of 2021. I also gave a talk at the group seminar on out-of-distribution generalization.
I attended for one month to the Les Houches summer school to learn some physics and mathematics of machine learning from world-class teachers and participants.
I gave a talk at Math Machine Learning seminar MPI MiS + UCLA! Here is the link.
I gave a virtual talk on the loss Landscape geometry of neural networks! Here is the link.