32-G484 (Stata Center)
32 Vassar Street
Cambridge, MA 02142
My current research interests are in developing methods for quantifying uncertainty and improving robustness in ML models. The ability to provide precise performance guarantees while handling challenging and realistic scenarios is critical to reliable, consequential-decision making. I also work on few-shot learning with pre-trained language models, as well as machine reading of unstructured text for automatic question answering, particularly at large scales.
Before coming to MIT, I worked as a research engineer at Facebook AI Research in New York (2015-17). I graduated with a degree in Mechanical Engineering from Princeton University in 2015, where my undergraduate thesis work was in robotics.
|May 8, 2021||Our work on few-shot conformal prediction is accepted to ICML 2021! We will also be presenting this work (and other conformal prediction work) as a spotlight at the DF-UQ workshop.|
|Mar 10, 2021||Our work on robust fact verification is accepted to NAACL 2021!|
|Jan 12, 2021||Our work on efficient large-scale conformal prediction is accepted to ICLR 2021!|
|Jan 1, 2021||New preprint on better few-shot prediction with pre-trained language models.|
|Feb 1, 2020||Completed an internship at Google Research with Kenton Lee. Checkout our EMNLP paper, CapWAP!|