Projects

Research project on studying the technologies for explaining the behaviors of machine learning models.

Convolutional Neural Networks training time prediction

This project aims to develop a mechanism that enables efficient development of artificial intelligence systems capable of identifying a wide variety of actions. By this mechanism, new services supporting human activities, such as monitoring, watching and assistance, are expected to be developed more efficiently.

A story behind the development of hananona, which is a flower classification system

This project aims to improve the efficiency of land-based salmon farming by utilizing artificial intelligence.

This project aims to develop a tool that automatically extracts raw data from figures and tables in materials science papers to utilize them in machine learning.

Fine-grained everyday human action recognition project, in which we research and develop an AI capable of understanding and question-answering about people's actions in home.

We advocate evidence-based performance tuning (EBT) for scientific applications.

This project focuses on fine-grained analyses of structural changes made between revisions of programs.