Skip to Content

Five questions with 2nd place winner of the 2020 Regeneron Science Talent Search, Jagdeep Bhatia

By Aparna Paul

Jagdeep Bhatia
Portrait of STS 2020 second place winner, Jagdeep Bhatia/ Illustration by Amy Wike Illustration by Amy Wike

Jagdeep Bhatia, the second-place winner and recipient of a $175,000 award in the Regeneron Science Talent Search 2020, is soon headed to MIT to deepen his knowledge of artificial intelligence and machine learning. These fields teach computers to learn without explicitly programming them; this has helped to make progress in diverse fields from banking to medicine. For example, some tasks that may be too time-consuming to rely wholly on people, such as scanning job resumes or loan applications can be more efficient by using AI. In medicine, for example, AI research has the potential to help a surgical robot adapt to the anatomy of a patient in real-time, in the operating room.

Though there are many pros to AI and machine learning, Jagdeep says that one serious limitation is that “machine learning models have been found to discriminate against certain people, for instance, when applied in our criminal justice system.” The other challenge Jagdeep points out is that teaching a computer seemingly straight forward tasks can be and often is data-expensive. “Even teaching a computer simple skills, like the ability to distinguish between images of cats and dogs, requires an unreasonably large amount of data.”

For this reason, Jagdeep has chosen to focus his work on a subfield of machine learning called interactive learning. He explains that this is “where a computer learns from a teacher (often a human) by asking questions and receiving feedback. The better the questions the computer asks, the better the feedback they receive from the teacher, and the faster they are able to learn. I came up with a way for a computer to ask just the right questions so they are able to learn the fastest, minimizing the amount of data needed.”

Jagdeep hopes that the mathematical techniques he used in his work could be applied to advance these fields and make algorithms less data-expensive and more robust.  One day, we’re sure to see Jagdeep lead AI breakthroughs in the world. For now, let’s learn more about him below!

Which sci-fi advancement are you upset we don’t have by now?

Full-dive VR technology. Right now, the capabilities of virtual experiences are limited to sight and sound, but what if we could also experience touch, smell and taste in the virtual world? This would really take Zoom meetings to the next level. Instead of just seeing and hearing people, you would get a full sense of spatial awareness! However, since this technology would likely be interfacing with peoples’ brains using electrical signals, serious ethical considerations are needed before it becomes mainstream.

Which scientist would you want to solve scientific mysteries with and why?

Entrepreneur and scientist, Elon Musk! Not only do I admire his spirited work ethic and am inspired by his unconventional way of problem-solving, but think that Elon’s projects are just so cool and futuristic. As Space X, Tesla and the Boring Company improve our infrastructure on Earth and in space, it’s not unreasonable that we will soon be able to establish lunar and martian colonies, manipulate the trajectories of celestial bodies and begin mining asteroids for rare elements. Achieving these feats will definitely be a giant leap forward for mankind, and it’s exciting to think that we will probably see this in action within our lifetime! 

What would you invent if you had all the resources in the world?

Imagine a robot that could learn to ride a bike without forgetting how to walk, that could design an original painting, that could quickly learn from its mistakes, that could apply prior knowledge to problem-solve, that could express sympathy. Empowered with true artificial intelligence, machines would be able to sustain conversation, explore places unsafe for humans like extraterrestrial landscapes and disaster zones, derive solutions to a multitude of unsolved mathematical and scientific problems, and (most importantly) wash your dishes.

What was the most surprising part of having STS is be a virtual event?

I don’t think I fully appreciated the level of organization that could go into holding a virtual event until I set up 4 boxes worth of tech equipment, was transported through a maze of Zoom breakout rooms by STS staff, and went searching for geocaches that STS had hidden in my town! It was also surprising (and frankly kind of cool) how one of the judging questions utilized the zoom meeting and would not have been answerable had we not been on a video call!

What was your most favorite or memorable experience from virtual STS? And why?

It was incredible to be part of such a positive, polite and passionate group of people! Whether we were playing skribbl and codenames or learning about each others’ talents through show-and-tell, I had a lot of fun with the other finalists. I think my most memorable experience was during orientation when we all opened our caricatures together!

Share your research for a chance to win incredible prizes and meet like-minded young leaders in science, just like Jagdeep, by applying to the Regeneron Science Talent Search 2021!
Application now open for the Class of 2021 through November 12, 2020.