*********************************
There is now a CONTENT FREEZE for Mercury while we switch to a new platform. It began on Friday, March 10 at 6pm and will end on Wednesday, March 15 at noon. No new content can be created during this time, but all material in the system as of the beginning of the freeze will be migrated to the new platform, including users and groups. Functionally the new site is identical to the old one. webteam@gatech.edu
*********************************
Ph.D. Thesis Proposal Announcement
Title: Human-Centered AI Understanding: Scalable, Interactive Visual Analytics of Machine Learning
Minsuk Brian Kahng
Computer Science PhD Student
School of Computational Science and Engineering
College of Computing
Georgia Institute of Technology
https://minsuk.com
Date: Monday, October 15th, 2018
Time: 1:00pm to 3:00pm (EDT)
Location: KACB 1116E
Committee:
----------------
Dr. Polo Chau (Advisor, School of Computational Science and Engineering, Georgia Institute of Technology)
Dr. Sham Navathe (School of Computer Science, Georgia Institute of Technology)
Dr. Alex Endert (School of Interactive Computing, Georgia Institute of Technology)
Dr. Martin Wattenberg (Senior Staff Research Scientist, Google)
Dr. Fernanda Viégas (Senior Staff Research Scientist, Google)
Abstract:
----------------
While machine learning has led to major breakthroughs in many domains, understanding machine learning models remains a fundamental challenge. They are often used as "black boxes," which could be detrimental. How can we help people understand complex machine learning models, so that they can learn them more easily, use them more effectively, and trust them more confidently?
My dissertation addresses these fundamental and practical challenges in the understanding of machine learning models through a human-centered approach, by creating novel visualization tools that are scalable, interactive, and easy to learn and to use. With such tools, users can better understand the underlying mechanisms of models, by visually exploring how the models process large datasets. Specifically, my dissertation focuses on three complementary parts:
(1) Conceptual understanding of models via interactive experimentation: designing interactive tools that broaden people's education access to learning complex deep learning models (e.g., GAN Lab);
(2) Visual analysis of models for industry-scale datasets: developing scalable visual analytics tools that help engineers interpret deep learning models through exploration of intermediate outputs (e.g., ActiVis deployed by Facebook); and
(3) Interactive exploration and discovery for actionable insights: designing interactive tools that support discovery of actionable insights through exploration of important data groups over different analytics stages, such as data preparation (ETable) and model selection (MLCube).
My research has made significant impact to society and industry. Our GAN Lab tool for understanding GAN training dynamics has been open-sourced, with its demo used by over 18,000 people from 119 countries. Our ActiVis system for deep learning visualization has been deployed on Facebook's machine learning platform.
----------------
