From Data to Decisions: On Learning, Prediction, and Action in the Open World

Eric Horvitz, Microsoft

A confluence of advances has led to an inflection in our ability to collect, store, and harness large amounts of data for generating insights and guiding decision making in the open world. Beyond study and refinement of principles, fielding real-world systems is critical for testing the sufficiency of algorithms and implications of assumptions-and exploring the human dimension of computational solutions and services. I will discuss several efforts pushing on the frontiers of machine learning and inference, highlighting key ideas in the context of projects in healthcare, transportation, and citizen science. Finally, I will describe directions with the composition of systems that draw upon a symphony of competencies and that operate over extended periods of time.

Speaker Biography

Eric Horvitz is a Distinguished Scientist at Microsoft Research. His interests span theoretical and practical challenges with developing systems that perceive, learn, and reason, with a focus on inference and decision making under uncertainty and limited resources. He has been elected a Fellow of the AAAI, the AAAS, and of the American Academy of Arts and Sciences. He received PhD and MD degrees at Stanford University. More information about his research, collaborations, and publications can be found at