Tue11Oct201612:00 pm - 01:00 pmNQ 3100 (The Ehrlicher Room), North Quad, 105 S. State St. Ann Arbor MI 48109
[Talk] Shiqing(Licia) He and Xin Rong
VIZITCARDS is a card-driven workshop developed for our graduate infovis class. The workshop is intended to provide practice with good design techniques and to simultaneously reinforce key concepts. VIZITCARDS relies on principles of collaborative-learning and research on parallel design to generate positive collaborations and high-quality designs. From our experience of simulating a realistic design scenario in a classroom setting, we find that our students were able to meet key learning objectives and their design performance improved during the class. In this presentation, we describe variants of the workshop, discussing which techniques we think match to which learning goals.
Shiqing(Licia) He is a PhD student at the UMSI. She is currently working with Prof. Eytan Adar. While focusing on information visualization research, Licia is also conducting projects and researches in design, game, and generative art. She hopes to find ways to merge traditional art with technology.
We present a novel programming IDE driven by a neural network embedding model. In the IDE, the end-users describe their goal in natural language, and the system proposes code solutions, or locates the part of the existing code that needs modification. The IDE includes a novel user interface, called nested-layer spotlight search, which allows rapid navigation of complex APIs. The backend is supported by a bimodal embedding network that jointly models natural language and programming language. The model is trained on data extracted from Stack Overflow, Github, programming textbooks, and various other online resources. Through lab and simulation studies, we demonstrate the utility and the accuracy of the system, especially in the context of supporting data scientists in creating visualizations. This is a practice talk for the upcoming UIST '16 conference in Tokyo, Japan.
Xin Rong is a doctoral candidate at the University of Michigan, School of Information. His advisor is Professor Eytan Adar. His research focuses on creating and evaluating intelligent systems that support task completion for end users and programmers driven by data mining and machine learning approaches. He also focuses on building visualization systems for interpreting and diagnosing deep neural networks. He received his Bachelor degree in automation from Tsinghua University in 2011, and has had internships at Baidu, Google, and Microsoft Research. He is expected to graduate in 2017 and is currently on the job market.