Faculty Recruiting Support CICS

NLP for Natural Documents

05 Mar
Thursday, 03/05/2020 11:45am to 1:15pm
Computer Science Building, Room 150 & 151
Machine Learning and Friends Lunch
Speaker: Amanda Stent

Today's finance industry is continuously searching for alpha, through more advanced modeling and through alternative sources of data. Many alternative sources of data are not raw numerical data but natural documents - human-readable documents containing tables, graphics and text. In this talk, I will present an overview of how we use NLP at Bloomberg to extract information from natural documents, and highlight some research challenges. I also present a case study of how such information can be combined with market data for better predictive modeling.