Title: Interpretable Machine Learning Based on Visual Analysis
Abstract: Interpretable machine learning aims to make the decision-making process of machine learning models more transparent to researchers and practitioners, so as to achieve effective communication and collaboration between human and machine. This report will introduce our visual analysis framework for machine learning models. This framework breaks out of the traditional visual analysis mechanism of “analyze first and then visualize” in a single-direction, and organically combines machine learning methods with interactive visualization methods, so as to better help users understand complex models and their outputs, analyze, diagnose, and constantly improve Machine Learning Models. It provides technical basis for users to select, utilize, and improve machine learning models. At last, combining with specific application examples, such as deep learning model and integrated learning model analysis, we introduce the visual analysis technology based on this framework.
Biography: Dr. Shixia Liu is a permanent associate professor at Tsinghua University. Her main research interests are Visual Analysis, Text Mining, and Information Visualization. She served as the Papers Chair of CCF Class A conference IEEE VIS (VAST) 2016 and 2017 (the first time for an Asian scholar), Associate editor-in-chief of IEEE Transactions on Visualization and Computer Graphics, and a former member of editorial board, Associate Editor of IEEE Transactions on Big Data and ACM Transactions on Interactive Intelligent Systems, and Chair of the Program Committee of IEEE Pacific Visualization 2015 of the International Visualization Conference. At the same time, she is a member of Editorial Board of the journal Information Visualization and is a member of the program committees of several international conferences, such as InfoVis, VAST, CHI, KDD, ACM Multimedia, ACM IUI, SDM and PacificVis. She served as Co-chair of IEEE VIS 2014 Meetup (IEEE VIS Organizing Committee) and Co-chair of IEEE VIS 2015 Tutorial (IEEE vis Organizing Committee). To learn more, please visit her profile page at:
http://cgcad.thss.tsinghua.edu.cn/shixia/.
Time: 14: 00-15: 30, Saturday, November 9, 2019
Venue: Conference Room 601, Administration Building
Organizer: School of Artificial Intelligence
