profile photo

Vivek Ramanujan

I am currently a predoctoral researcher on the PRIOR (vision) group at the Allen Institute for Artificial Intelligence (AI2), where I work on computer vision and machine learning. I am advised by Mohammad Rastegari and Aniruddha Kembhavi. I also collaborate with Ali Farhadi at the University of Washington Seattle.

Email  /  CV  /  Github

Research

I'm broadly interested in computer vision, machine learning, and optimization.

clean-usnob Supermasks in Superposition
Mitchell Wortsman*, Vivek Ramanujan*, Rosanne Liu, Aniruddha Kembhavi, Mohammad Rastegari, Jason Yosinski, Ali Farhadi
preprint, 2020

We present an application of hidden networks for continual learning, capable of learning thousands of tasks without catastrophic forgetting. We solve tasks individually, each solution corresponding to a subnetwork of a randomly initialized neural network. Using a superposition of these subnetworks, we demonstrate that the viability of this model for task inference. Finally, we introduce a coherent hierarchy for continual learning problems.

Code / Blog

clean-usnob Soft Threshold Weight Reparameterization for Learnable Sparsity
Aditya Kusupati, Vivek Ramanujan, Raghav Somani, Mitchell Wortsman, Prateek Jain, Sham Kakade, Ali Farhadi
To appear at the International Conference on Machine Learning 2020

We introduce a new strategy for pruning neural networks based off of the soft threshold reparametrization technique from signal processing. The layerwise sparsity budgets allow for very sparse but still highly performant trained models across a variety of architectures and tasks.

clean-usnob What's Hidden in a Randomly Weighted Neural Network?
Vivek Ramanujan*, Mitchell Wortsman*, Aniruddha Kembhavi, Ali Farhadi, Mohammad Rastegari
Computer Vision and Pattern Recognition 2020

We demonstrate that you can find untrained subnetworks of common overparametrized convolutional neural networks at initialization that achieve performance similar to their densely trained counterparts.

Code and Project Page

clean-usnob Improving Shape Deformation in Unsupervised Image-to-Image Translation
Aaron Gokaslan, Vivek Ramanujan, Kwang-In Kim, Daniel Ritchie, James Tompkin
European Conference for Computer Vision, 2018

We improve on CycleGAN by allowing for better shape deformation between more disparate domains.

Service
Brown CS Teaching Assistant, Computer Vision CS146 Spring 2018

Teaching Assistant, Machine Learning CS142 Spring 2018

Teaching Assistant, Applied Artificial Intelligence CS141 Spring 2017

Teaching Assistant, Deep Learning CS2951K, Fall 2016

This website's source is slightly modified from Jonathan Barron's website