Yoichi Ochiai

Design and Deployment of a xDiversity AI platform for Audio-Visual-Tactile Communication towards an inclusive society

Research Director

Yoichi Ochiai

Collaborator

Ken Endo	Researcher Sony Computer Science Laboratories, Inc.
Yusuke Sugano	Associate Professor Graduate School of Information Science and Technology Osaka University
Tatsuya Honda	Marketing Strategy Headquarters FUJITSU LIMITED

Outline

This project aims to explore AI-assisted human-machine integration techniques for overcoming impairments and disabilities. By connecting assistive hardware and auditory/visual/tactile sensors and actuators with a user-adaptive and interactive learning framework, we propose and develop a proof of concept of our “xDiversity AI platform” to meet the various abilities, needs, and demands in our society. Our final goal is a social design and deployment of the assistive technologies towards an inclusive society.

Tokuyasu Kakuta

Drafting support for legal texts with AI technology

Research Director

Tokuyasu Kakuta

Outline

The purpose of this research is, with AI technology, to support drafting legally effective documents such as laws and contracts called “legal texts”. In this research we develop an AI system with which users can draft legal texts by inputting the parts of templates of the texts and by which these templates are composed automatically corresponding to various cases from a large number of legal texts. Furthermore using the templates we will develop an automatic verification system for legal texts and an educational supporting system to draft legal texts.

Yuji Sekiya

Real-time threat detection and prediction by analyzing cyber threat big data

Research Director

Yuji Sekiya

Collaborator

Yuki Kadobayashi	Professor Division of Information Science Nara Institute of Science and Technology
Keiichi Shima	Deputy Director Research Institute IIJ Innovation Institute
Satoshi Matsuura	Associate Professor Global Scientific Information and Computing Center Tokyo Institute of Technology

Outline

This research aims to counterwork complex cyber attacks and predict the types of cyber threats and its impacts on managed systems by real-time analysis of cyber threat big data. The counterplan of cyber attacks is critically dependent on the persons who have the knowledge of cyber threats. In case of the organization which doesn’t have such a person, the counteraction against the cyber incidents is delayed, and it increases the problem. We propose the decision support system to help the incident response in the organization. Our contributions in this research are (1) proposing the common analysis methods of cyber threats big data, (2) providing the methodologies and algorithms for the analysis, and (3) publishing the vectorized data as open data.

Tanaka Toshihisa

Development of AI-based diagnosis and treatment-aid for epilepsy by automated interpretation of electroencephalogram

Research Director

Tanaka Toshihisa

Collaborator

Hidenori Sugano

Associate Professor Department of Neurosurgery Juntendo University

Outline

This project establishes “AI-based diagnosis and treatment-aid” system that learn electroencephalograms (EEG) of epilepsy patients and diagnoses by clinicians. Japan has about one million patients of epilepsy, but there are very limited numbers (about 600) of clinicians who can read and interpret EEG of patients. Epileptic seizures sometimes cause serious traffic accidents, and thus it is highly desirable to take a social measure. In this project, we aim to create an AI that can learn the diagnosis by medical specialists in epilepsy for EEG data measured in hospitals. To do that, we develop techniques to construct a dataset of EEG and to learn the interpretation of the data.

Hiroki Matsutani

An Edge Learning Infrastructure Supporting Realtime and All-Data Capabilities

Research Director

Hiroki Matsutani

Collaborator

Masaaki Kondo	Associate Professor Graduate School of Information Science and Technology The University of Tokyo
Yasuhiko Shioda	Executive Officer Solution Division Fixstars Corporation

Outline

A huge amount of stream data is generated continuously by industries and network services. For automated monitoring and anomaly detection by learning such big data, tendency changes of targets should be reflected to the learned parameters immediately, though deep learning is a time-consuming task. In this project, we will develop an AI infrastructure that can reflect recent data tendency in addition to all the past data to the learned parameters. We will validate our concept through experiments targeting industries and network services.

Ken’ichi Morooka

Innovative Computer-Aided Cancer Diagnosis with Artificial Intelligence using 3D Shapes and Appearance of Cells

Research Director

Ken’ichi Morooka

Collaborator

Eiji Ohno	Professor Faculty of Health Sciences Kyoto Tachibana University
Hajime Nagahara	Professor Institute for Datability Science Osaka University

Outline

Cytological diagnosis of cancer using specimen is useful to detect cancers at an early and easily curable stage. The project aim is to construct an innovative computer-aided cervical cytology with artificial intelligence for detecting cancer cells reliably. To achieve this, the project will develop technologies for 1) recovering 3D shapes of cells from multi-focus super high-resolution images and 2) identifying normal and cancer cells by deep neural networks using 3D shapes and appearance of cells.