Yali Bian, Ph.D. student in the Department of Computer Science
Graphic from the paper “DeepVA: Bridging Cognition and Computation through Semantic Interaction and Deep Learning” 

According to Yali Bian, the Sanghani Center’s proclivity for encouraging interdisciplinary research is an added benefit while working on his dissertation topic, “Interactive Deep Learning for Semantic Interaction.” It encompasses several different research areas like human computer interaction, deep learning, visual analytics, and explainable AI. 

Bian is exploring ways to provide user-friendly interactive visualization systems to users unfamiliar with deep learning so that they can make full usage of powerful deep learning models.

A good example, he said, is building a truly hybrid human-AI co-learning system than can assist intelligence analysts in their sensemaking tasks. The analysts could gain useful feedback from powerful deep learning models without knowing how to manipulate the model.  

“My interest in this particular topic stems from wanting to address what I think of as a ‘last mile problem,’’’ said Bain, who is advised by Chris North. “I would like to see powerful machine learning models fully accessible to laypeople – to the extent that they could design a personal model.”

In April, his paper, “DeepSI: Interactive Deep Learning for Semantic Interaction,” will be included in proceedings at the ACM IUI2022 conference. The study proposes framework that can be fine-tuned to integrate deep learning into the human-in-the-loop interactive sensemaking pipeline.

Other papers published while Bian has been at the Sanghani Center include the following collaborations with North and other Ph.D. students: “DeepVA: Bridging Cognition and Computation through Semantic Interaction and Deep Learning,” in the proceedings of the IEEE VIS 2019 Workshop on Machine Learning from User Interactions for Visualization and Analytics; and “Evaluating Semantic Interaction on Word Embeddings via Simulation,” presented at the EValuation of Interactive VisuAl Machine Learning Systems Workshop at the same conference.

Bian holds a master’s degree in computer science from Zhejiang University, China, and – along his path to a Ph.D. – earned a second master’s in computer science from Virginia Tech.

“I came to Virginia Tech in 2016 and I think that Blacksburg provides an ideal atmosphere,” said Bian. “There are a lot of great hiking trails around campus and it is a great place to focus on study and research so it helps you find a good academic work-life balance.”

Bian is planning to receive his Ph.D. in 2021 and would like to work in a research industry position after graduation.