to Understand Natural Language Queries Abstract: Question answering (QA) requires computers to understand and act upon queries from humans. The process of transforming natural language into precise and executable meaning representations is known Thursday, April 19, 2018 as semantic parsing. Can we induce a semantic parser from data using machine learning? The challenge lies in blending the symbolic and 4:00PM - 5:00PM statistical sides of AI while retaining the accuracy levels required in finance. Hewlett Room, 32-G882 Ray and Maria Stata Center In this talk, we will introduce our research into learning grammars for financial QA in collaboration with Prof. Yoav Artzi of Cornell University. Cambridge, MA 02139 We utilize the combinatory categorial grammar (CCG) framework, which aligns syntax (functions of parts of speech) and semantics (typed lambda calculus). We will assess the challenges of learning and Please RSVP to: inference in this approach and compare it to other semantic parsing methods. http://bit.ly/BloombergTechTalkSpring2018 Presenter: Gautam Shine, ML Research Scientist, Bloomberg Questions? Contact: Gautam Shine is a researcher and engineer in the Machine Learning Philip Arsenault at pmarsena@csail.mit.edu group at Bloomberg. He specializes in natural language understanding and the development of grammars and parsing algorithms for question answering systems. Prior to joining Bloomberg in 2017, he earned his Ph.D. in electrical engineering at Stanford University.