My goal is to enable reliable communication in natural langauge between machines and humans. Recent research directions include: (1) Text generation: How do we ensure that the generated text are not just coheret, but also factually correct? Beyond factuality, how do we generate novel, creative text in a controllable way? (2) Robust language understanding: Statistical machine learning systems suffer from spurious correlations in the data, resulting in biased prediction and catastrophic errors. How can we learn models that are robust to dataset biases? (3) Dialogue systems: Building an effective dialogue agent requires understanding in a broad sense, e.g., how to reason about structured knowledge and learn efficient strategies for a specific task?
Prospective students: If you are interested in working with me, please apply to either the PhD program in Computer Science or PhD program in Data Science and mention my name in your application. If you are a student at NYU, please drop me an email.