Bloomberg Info Session
Talk Title: How to Train Your Search Engine
Speaker: Jon Dorando, Applied ML Scientist
Date: January 13, 2015
Time: 5:00pm - 6:00pm
Location: Gates Building, room 104
Dinner will be served!
Please RSVP through Handshake at https://stanford.joinhandshake.com/events/10153
Stanford Students only and a Stanford ID will be required at check-in
Abstract:
Relevance ranking of search engine results is one of the most important problems in Information Retrieval. The financial markets are a rich and complex domain for search and have pose unique challenges for ranking. The sources are many and disparate, like domains with rich structured data such as company and security attributes, textual data like frequently asked questions and analysis reports as well as timely data like news. Not only is the domain complicated, but some of the techniques that work for web search have to be adapted and reconsidered in an enterprise context.
In this talk, we will discuss our approach to federated search at Bloomberg. We will start with how search at Bloomberg compares to web search and discuss challenges that are unique to us. This will be followed by a discussion of how we used Machine Learning techniques to design, evaluate and maintain our ranking models.
