Chenglong Wang - Microsoft Research
Jan. 14, 2022, 2:30 p.m. - Jan. 14, 2022, 3:30 p.m.
Virtual (see link below)
Hosted by: Xujie Si
Abstract: Data scientists manipulate and visualize data to gain insights. However, while experienced data scientists can solve these problems and achieve efficient and flexible analysis using programming languages like R, inexperienced users struggle to achieve similar results using only interactive tools.
We developed a synthesis-powered visualization tool, Falx, to help end-users traverse this gap. In Falx, users demonstrate example mappings from input data values to visual channels, and Falx automatically transforms and visualizes the full dataset. Falx's magic comes from (1) a novel interaction model evolved from the grammar of graphics that allows users to express complex tasks with less effort, and (2) a scalable synthesis algorithm that leverages abstract program reasoning to efficiently solve the combinatorial program search problem. The result: Falx can efficiently solve 75% of 83 practical tasks within 10 seconds, and 33 users in our study can confidently use Falx to solve challenging visualization tasks they cannot easily solve otherwise. I will conclude with our observations of programmability gaps for end-users in the field, and how PL-HCI-AI research breakthroughs could vastly expand access to the power of programming to bridge these gaps.
Bio: Chenglong Wang is a computer science researcher in Microsoft Research deep learning group. His research focuses on building novel synthesis-powered tools to democratize data analysis. Prior to that, Chenglong finished his Ph.D. from the University of Washington working with Ras Bodik and Alvin Cheung. Chenglong was lucky to receive a few awards from the research community: SIGMOD'17 best demo award, InfoVis'18 best paper award, and CHI'21 best paper award.
Zoom link: https://mcgill.zoom.us/j/84280715850
Virtual reception after the talk in Gather: https://gather.town/app/tYHHMh7tPcPw9037/reception