Ananye Agarwal

Email: ananyea [at] andrew [dot] cmu [dot] edu

I am an incoming PhD student at Carnegie Mellon University's Machine Learning Department.

I recently graduated from IIT Delhi with a B.Tech. in Computer Science and Engineering. In the past, I have been lucky to work with Prof. Mausam on neuro-symbolic AI and with Prof. Manik Varma on Extreme Classification. I have also had the good fortune of interning at Microsoft Research India where I worked Dr. Ankit Garg on algebraic complexity.

profile photo

Email  |  Twitter  |  Github

Research

My goal is to build intelligent agents that can efficiently generalize to solve a diverse set of tasks. To this end, I am interested in techniques from reinforcement learning, machine learning, computer vision and cognitive psychology. In the past, I have been interested in Extreme Classification and neuro-symbolic AI.

End-to-End Neuro-Symbolic Architecture for Image-to-Image Reasoning Tasks

Ananye Agarwal, Pradeep Shenoy, Mausam
arXiv 2021
arXiv

We describe a novel neuro-symbolic-neuro architecture that is able to solve tasks where both the input and ground truth output are images.
SiameseXML: Siamese networks meet extreme classifiers with 100M labels

K. Dahiya, A. Agarwal, D. Saini, K. Gururaj, J. Jiao, A. Singh, S. Agarwal, P. Kar and M. Varma
ICML 2021
PDF | code

We show how SiameseXML can be combined with extreme classifiers to give SOTA while being massively scalable.
Website template modified from Jon Barron's repo.