Akarsh Kumar

alt text 

Welcome!
I want to understand intelligence and all the processes that emerge it.

Contact akarshkumar0101(at)gmail(dot)com

Research Interests

Broadly, I want to create+analyze artificial intelligence (AI) agents that reason+learn more efficiently+robustly in increasingly complex environments.
Specifically, I am interested in neural networks (NNs), reinforcement learning (RL), and meta-learning combined in the framework of AI Generating Algorithms (AIGAs).
Technically, I want to implement these ideas by combining stochastic gradient descent (SGD) with quality-diverse and open-ended evolutionary computation.

Bio

I will be joining MIT CSAIL in Fall 2022 to do a Ph.D. in AI!
I'll be working with Phillip Isola on AI-GAs (meta-learning) and Open-Endedness (evolutionary computation) for creating intelligent embodied agents!
My research will be funded by the NSF GRFP!

I recently graduated from UT Austin with a B.S. in Electrical/Computer Engineering.
Senior year, I did NN architecture research in VITA with Prof. Atlas Wang.
Senior year summer, I did neuroevolution and RL research in LARG with Prof. Peter Stone and Prof. Risto Miikkulainen.
Sophomore and junior year, I worked with Prof. Alexander Huth on visually estimating contact forces in hand-object interaction.
Before that, I went to highschool in the Arkansas School for Mathematics, Sciences, and the Arts.

Here is my CV.