Időpont: október 17, 17h

Helyszín: BME Q Épület, QBF08 terem

Daniel Sturman (Cloudera, USA): Interpretable machine learning products

Machine learning algorithms are increasingly making decisions for and about people in modern society. Algorithms decide which e-mails reach our inboxes, whether we're approved for credit, and whom we get the opportunity to date. An interpretable algorithm is one whose decisions you can explain to your clients and to your regulators as well. You can better rely on such a model to be safe, accurate and useful since you understand it. An accurate model that is also interpretable can offer insights that can be used to change real-world outcomes for the better. Interpretable machine learning will allow for a more collaborative relationship between machines and humans and will help to resolve ethical concerns related to the use of ML.

The presentation will cover how interpretability can make deep neural networks models easier to understand, and describes LIME, an open source tool that can be used to explore what machine learning classifiers (or models) are doing.

Short biography:

Daniel Sturman is Sr Vice President, Engineering & Support at Cloudera. Prior to Cloudera Dan was at Google where he led development of cloud products including Google Compute Engine, Google App Engine, and Kubernetes as well as the internal cluster management systems that manage all computation across Google’s fleet of servers. Daniel was also closely involved in the growth of Google’s New York office, leading a team working on a range of challenges faced by Google’s internal software infrastructure. Previous to Google, Daniel was director of development for DB2 on Linux, Unix and Windows in IBM's Information Management division. Daniel started at IBM as a researcher at the IBM T. J. Watson Research Center, where his research focused on revolutionizing the way people build and use distributed systems. His research concentrated on technologies for enterprise messaging and cluster computing. He holds a Ph.D. and master's degree in computer science from the University of Illinois at Urbana-Champaign and a bachelor's degree in computer science from Cornell University.