
Victor S. Bursztyn
PhD Candidate @ Northwestern University
2
presentations
SHORT BIO
I am a PhD candidate at Northwestern University advised by Prof. Larry Birnbaum. I am generally interested in the computational modeling of conversations; more particularly, I am interested in how large language models can help solve "pragmatics gaps" that are especially evident in conversations geared towards recommendations. One example is the gap between a critique that a user utters when given an imperfect recommendation (e.g., "This restaurant doesn't look good for a date") and the preference that should be inferred (e.g., "I prefer a more romantic place") such that a better recommendation can be made—you may find more information in our short-paper at the conference. Overall, I see this as an applied space where common sense reasoning can be studied and put in practice. Outside of academia, I worked as a research scientist at Dell EMC's Brazil R&D Center for 1.5 years (with 7 patent applications filed), and I was a research science intern at Adobe during the Summer of 2021 (with 1 patent application filed).
Presentations

It doesn't look good for a date: Transforming Critiques into Preferences for Conversational Recommendation Systems
Victor S. Bursztyn and 5 other authors

"It doesn't look good for a date": Transforming Critiques into Preferences for Conversational Recommendation Systems
Victor S. Bursztyn and 5 other authors