Meet ML@GT: Abhishek Das Wants to Stop Climate Change and Develop AI Agents with Human-Level Skillsets

Abhishek Das (1)

The Machine Learning Center at Georgia Tech (ML@GT) is home to many talented students from across campus, representing all six of Georgia Tech’s colleges and the Georgia Tech Research Institute (GTRI).

These students have diverse backgrounds and a wide variety of interests both inside and outside of the classroom. Today, we’d like you to meet Abhishek Das, a fourth year computer science Ph.D. student who is expecting to graduate in May 2020. Inspired by many, and an avid scuba diver, reader, and podcast listener, Das hopes to continue his work on developing artificial agents with human-level skillsets and affecting climate change after graduation.

Advisor: Dhruv Batra

Hometown: New Delhi, India

Current Georgia Tech degree program/year: 4th year computer science Ph.D. student. I’m expecting to graduate in May 2020.

Other degrees earned and from what institution: Bachelor’s in Electrical Engineering from Indian Institute of Technology Roorkee

Tell us about your research interests.

My research focuses on building artificial agents that can see, talk, and act the way we do as humans. This has wide-ranging applications — from assistive chatbots for the visually impaired, to educational tools for children, to natural language interaction with augmented reality (AR) and virtual reality (VR) headsets and in-home mobile robots.

What is it about machine learning that excites you?

It feels magical when it works! More seriously, the underlying building blocks and algorithms used in machine learning today are shared across unrelated domains, which massively compounds the impact of every breakthrough. For instance, the same model class that can recognize dogs in images can also predict properties of chemical compounds, with minimal domain-specific engineering.

What are some of your hobbies?

I’ve been playing the piano since I was 5. Other than that, I play soccer, enjoy taking photos, read quite a bit and occasionally scuba dive.

We just entered a new decade. What are you most looking forward to in the next 10 years?

Making progress on the climate crisis! That is likely the single biggest challenge for humanity in the near future. We desperately need technological innovations to 1) guide policies and resource allocation, 2) accelerate scientific breakthroughs towards carbon negative, and 3) reduce the psychological distance to the climate crisis in our daily lives.

Why did you choose Georgia Tech as the place to earn your Ph.D.?

Georgia Tech is an amazing school, but to be honest, it was incidental for me. I started my Ph.D. at Virginia Tech, and my advisor, Dhruv Batra, moved to Georgia Tech a semester after I started. I moved with him and have been glad that I made the transition.

What inspired or motivated you to earn a Ph.D. in machine learning?

I’d worked with Dhruv over the course of an internship before starting as a Ph.D. student. I loved what I was reading and working on during the internship, found research overall intellectually stimulating, and so decided to continue as a Ph.D. student.

What do you hope to do after graduation?

In the near future, I’ll likely continue research in the industry, and hopefully make a foundational dent in developing artificial agents with human-level skillsets, and the climate crisis.

What is your proudest accomplishment?

Professionally, there have been many ‑ my first publication, the first media coverage of one of my papers, the first 100 citations, my first talk at a conference, having ridiculously smart people around during my internships at DeepMind and FAIR — it’s all been quite surreal!

From my personal life, I recently got certified as an open water diver; that felt like quite an achievement (though is probably no big deal for most people!)

Who is someone that inspires you and why?

Many along various dimensions:

John Platt and Elon Musk — for the ambitious problems they’re going after.

Joelle Pineau and Devi Parikh — for being time management exemplars.

Julia Galef — for her clarity of thought (highly recommend listening to her Rationally Speaking podcast.)

Dhruv Batra — for constantly questioning authority and the status quo.

Ross Girshick — for the thoroughness and perfection of all his research.

Malika Favre, Gal Shir and Zaha Hadid — they are by far the most creative people I know of.

Podcast, movie, book, or tv show? Why is that your favorite medium and what is your favorite of your chosen medium?

Books > podcasts >>> tv shows > movies.

Some of my favorite books are “The Elephant in the Brain” by Kevin Simler and Robin Hanson, “The Martian” by Andy Weir, “The Uninhabitable Earth” by David Wallace-Wells, “Justice” by Michael Sandel, and “The Three-Body Problem” series by Liu Cixin.

I also love podcasts. Some favorites include Philosopher’s Zone by ABC Radio, Rationally Speaking by Julia Galef, More Perfect by Radiolab, Anthropocene Reviewed by WNYC Studios.

I don’t watch too many movies, but I love “Inception,” “Joker,” the Dark Knight series, and a Bollywood film called “3 Idiots.” My go-to TV shows are “Breaking Bad,” “The Crown,” “Narcos,” “Sacred Games,” “Friends,” and “Chernobyl.”

What is the most random app on your phone?

Chwazi — it’s a “finger chooser” app to decide teams or who goes first in a group. 🙂

What is your favorite thing about living in Atlanta? What are some of your go-to spots for studying or hanging out?

I really like the new Coda building. It has great design, amazing views of the city, and functional coffee machines.

Other than that, I sometimes hang out at Dancing Goats (Midtown) or Land of a Thousand Hills (Atlantic Station). They have good coffee. I also highly recommend the eggs benedict at Cafe Intermezzo in Midtown. It’s delicious!

 

Press Contact:

Allie McFadden

Communications Officer

allie.mcfadden@cc.gatech.edu

1 thought on “Meet ML@GT: Abhishek Das Wants to Stop Climate Change and Develop AI Agents with Human-Level Skillsets”

Leave a comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.