CS Talk - Wei Xu
Title: Multiple Instance Learning from Unlimited Text
Host: Daniel Spielman
Coffee/tea - 10:15, BCT MC035
The advent of big data provides both challenges and opportunities for processing natural language. My research focuses on scalable machine learning methods to understand the seemingly unlimited number of expressions in human language. In this talk, I will present a multiple instance learning model that learns variant expressions of the same meaning from Twitter’s massive data stream. This is one of the first models that can jointly reason about relations between words and sentences. I will highlight its value and broad applications in natural language generation, digital humanities and computational social science. I will also hint at how similar models can leverage and expand large knowledge bases.
Wei Xu is a postdoctoral researcher in the Computer and Information Science Department at the University of Pennsylvania. Her research interests lie at the intersection of machine learning, natural language processing, and social media. She has received the 5-year MacCracken Fellowship and completed her PhD in 2014 from New York University. During her PhD, she visited University of Washington for two years. She is organizing the ACL Workshop on Noisy User-generated Text, serving as the publicity chair for NAACL 2016 and as an area chair for EMNLP 2016.