Livestream:

https://northwestern.hosted.panopto.com/Panopto/Pages/Viewer.aspx?id=a65af02c-1aba-4f47-a1d7-aca901517ca2

Title:

Interactive Machine Teaching: Concepts and Lessons

Abstract:

Machine Learning (ML) can infuse systems with behaviors that rival or surpass human capabilities. This promise of creating systems that can empower people and organizations is met with the friction that creating ML models remains a complex task beyond the reach of non-ML experts. In this talk, I will describe Interactive Machine Teaching and its potential to simplify the creation of ML models. One of the key characteristics of IMT is its iterative process in which the human-in-the-loop takes the role of a teacher teaching a machine how to perform a task, and that the teacher does not need expertise about the underlying ML learning algorithm. After this introduction, I will summarize our insights and lessons from our research in this field, and what promising research directions lie ahead.

Biography:

Gonzalo Ramos is a Principal Researcher at Microsoft Research at Redmond, working on lowering the barrier of entry for people to teach to machines to perform useful tasks. He received his Ph.D. from the University of Toronto's Computer Science Department, specializing in HCI, with a focus on direct interactions. He worked as a Senior Design Technologist and UX Scientist at Amazon, and as a Scientist at Microsoft. His website can be found at https://www.microsoft.com/en-us/research/people/goramos/