As Rishi Bommasani, a recent graduate from the computer science master's program ('20), and a CS alumnus (B.A. '19), heads to Stanford University to pursue doctoral studies in the Natural Language Processing group, CS News caught up with him. While at Cornell, Rishi received the Computer Science Prize for Academic Excellence and Leadership as well as several Outstanding Teaching Assistant Awards, and helped establish the ongoing forum CS Research Night. At Stanford, Rishi's studies will be funded by a National Science Foundation (NSF) Graduate Research Fellowship.
CS News: How did you get involved in research at Cornell CS, specifically NLP?
Rishi: When I was a freshman, I met Professor Claire Cardie in her capacity as Director of Undergraduate Studies. I (somewhat naively) assumed she would be an all-knowledgeable resource on various courses in the department, for example, knowing the importance of prerequisites, etc. (I would refer undergraduates with these kinds of questions to ask Nicole Roy or Ryan Marchenese instead!) We met a few times and at some point she mentioned her research was natural language processing (NLP). I was intrigued by this and came to discover that some of the sentiment analysis algorithms I designed in high school actually belong to the broader field of NLP (and that sentiment analysis was pioneered by her longtime colleague and next-door office neighbor, Professor Lillian Lee). She recommended that I read Chris Manning's classic NLP textbook and eventually we arranged for me to do research over the summer after my freshman year. I have worked with Claire ever since on a variety of research problems in NLP.
CS News: What do you find compelling about Natural Language Processing (NLP)?
Rishi: Language is, in my mind, humanity's most remarkable innate gift. From a biological perspective, language, in all of its richness and complexity of expression, is a key differentiator in the communicative abilities of humans from other organisms. Simultaneously, mathematics is humanity's greatest achievement in formalizing knowledge and computation is the most recent manifestation of that, dramatically changing the world every day. At the intersection of language, mathematics, and computation lies NLP. For me, each of these domains are independently exciting and NLP is even more stimulating as there are many open problems that we have only scratched the surface of. From a pragmatic standpoint, NLP is also a field where progress on fascinating intellectual problems can have a meaningful impact on the world.
CS News: Why have you been interested in making research more accessible, especially at the undergraduate level?
Rishi: I have been very fortunate to be advised by Claire for several years and have reaped the benefits of her excellent guidance. It seems only natural to want to "pay it forward." However, there is far more to say on this subject. Many of the crucial open problems in computing science will require significant advances in both fundamental and applied research for us to see progress and solutions. Furthermore, by the nature of most academics and faculty, research can appear to an outsider as intimidating, tedious, and challenging. These are not unfair characterizations. One of my goals, whether it is in our CS Research Night series that has emerged as the premier venue for Cornell undergraduates to be exposed to CS research or in directly advising undergraduate researchers in NLP, has been to show that research can be welcoming, exciting, and rewarding. From a practical perspective, the field (and especially AI) has experienced an increased loss of talent to industry in the past decade or so. This so-called "brain drain" means we need to reinvigorate efforts to allow undergraduates to see what research holds in store and further push the envelope in terms of diversity and inclusion in research. While historically the start of a Ph.D. has been viewed as the key starting point of a journey in research, these circumstances have changed, and I believe heightened focus needs to be place on nurturing researchers at the undergraduate level as a result.
CS News: Congratulations on your acceptance at Stanford and your funding from the NSF.
Rishi: Thank you. I will start my Ph.D. in Fall 2020 in the Comptuer Science department at Stanford and will be affiliated with both the Stanford NLP group and the Stanford Artificial Intelligence Laboratory (SAIL). As you note, my graduate studies will be funded by a National Science Foundation Graduate Research Fellowship.
The Cornell CS community wishes Rishi the very best in his continued computer science studies.
Read what CS Professors Claire Cardie and Robert Kleinberg have to say about their former student (scroll to the second story).
Read Rishi's master's thesis, "Generalized Optimal Linear Orders" (2020), and find a biographical sketch (on page iii).