Interaction × A.I.
- JST CREST HCI for Machine Learning Symposium -
6F Hilobby, IREF bldg., Yayoi-campus, Hongo-district, The University of Tokyo [access]
Data driven intelligent information processing technologies such as machine learning are making rapid progress. In the JST CREST HCI for Machine Learning project, we are developing interaction technologies to support users of such data driven type intelligent information processing systems in data generation and learning phases as well as in execution phases. At this symposium, we introduce the results from our group so far and invite involved researchers to have discussions. We are waiting for active participation of people interested.
Program
13:00-13:15 Project Introduction. Interaction for Machine learning. Takeo Igarashi (Univ. Tokyo).
13:15-14:00 Invited Talk. Takayuki Ito.
14:00-15:00 Short presentations (Poster previews).
15:00-18:00 Poster presentations.
Registration
Here（Not required, but appreciated.）
Invited Talk
Research trend on visualization for machine learning / Takayuki Ito
Abstract: Machine learning is a recent emerging hot topic in visualization research communities. Many researchers try to "open the black box" by visually representing the behavior of training processes, results of model constructions, contents of training and test datasets of machine learning techniques. This talk firstly introduces recent trends on visualization for machine learning in recent visualization conferences and then presents the speaker's recent study on visualization assisting machine learning tasks.
Bio: Takayuki Itoh is a full professor of the department of information sciences of Ochanomizu University. He has received his B.S., M.S., and Ph.D. degrees from Waseda University in 1990, 1992, and 1997, respectively. He has been a researcher at Tokyo Research Laboratory of IBM Japan from 1992 to 2005. He has been an associate professor in Ochanomizu University since 2005 and a full professor since 2011. His research interest includes visualization, graph drawing, computer graphics, multimedia, and human-computer interaction.
Poster Presentations
Interaction
- Bayesian Information Gain for Interactive Level Control of Video Storyboard
- Interactive Design Exploration of Game Stages Using Adjustable Synthetic Testers
- Data-centric interaction for data transformation with Programming-by-Example
- Interactive Music Composition Tools for Deep Generative Models
Graphics
- Human-in-the-loop differential subspace search in high dimensional latent space
- 3D Model Synthesis from one single image for Anime Characters
- Data-driven Prediction of Aerodynamic Properties of 3D Objects
- Texture generation from sketch image
- Makeup system for transgender individuals
- Virtual Try-on by neural image translation
Machine Learning
- Low rank max margin majority voting
- Interactive Reinforcement Learning
- Integrated Development Environment with Interactive Scatter Plot for Examining Statistical Modeling
- Neural Network Programming with Live Visualization
Medical Imaging
- Tumor segmented database synthesis with human interaction
- Prediction of brain tumor using radiomics
- 3D Intracranical anuerysm dataset
- The surgical support using mixed reality by integrating virtual space and real space in neurosurgery
Contact
Igarashi lab. The University of Tokyo. igarashi.lab＠gmail.com