Ron Shprints

I'm a senior at MIT studying mathematics and computer science. My goal is to develop computational models that can understand the physical world beyond human-level intelligence in a scalable and principled manner. To do so, I aim to leverage techniques in unsupervised learning to tackle unsolved problems such as long-context modeling and learning under uncertainty. My research philosphy is guided by the following principles (if I really had to choose just 3 of them):
I don't limit myself to specific methodologies, but seek the mathematical foundations required to solve problems rigorously. I especially enjoy thinking about statistics.
Gaining empirical insights into deep learning models and their behavior is a great way to navigate research ideas.
Some of my best Eureka moments were inspired by methods and challenges from the natural sciences. I'd consider it "mission accomplished" if deep learning models could do real science one day.