News Releases

Toshiba and RIKEN to Collaborate in AI Research and Development

New Collaborative R&D Centre to advance self-learning AI
10 Mar, 2017

TOKYO—Toshiba Corporation (TOKYO: 6502) today announced that it has agreed with RIKEN, one of a Japan’s government funded research institutes, to work together to promote advances in fundamental technologies for AI. The collaborative work will be centered at a new research and development center, RIKEN AIP*1 Toshiba Collaboration Centre (RATCC), at the RIKEN Centre for AIP Project in Tokyo. By bringing together their leading-edge AI technologies, plus know-how and data Toshiba has built up through years of experience as a leading industrial company in sectors ranging semiconductors to infrastructure, the partners aim to achieve “Autonomous learning AI for innovative productivity”.

Toshiba R&D in AI focuses on three core areas:
ž ・  “Improving Plant Productivity” covers AI technologies applied at process improvement 
     in manufacturing facilities.
 ・   “Improving Intellectual Productivity” utilizes media data, to extract intelligence from
     data sources such as voices, images and texts.*2
 ・   “Automated Mobility and Robotics”, which enables automation in infrastructure inspection
     and maintenance, and carrying task in logistics and distribution.

Conventionally, AI requires data input by human to support learning, or a designed learning process, which are insufficient for coping with complicated problems. By promoting R&D in key areas in collaboration with RATCC, Toshiba aims to evolve beyond AI that requires human help and promote a shift to autonomous-learning AI.

In carrying out research and development at RATCC, Toshiba will draw on a broad range of cutting-edge technologies and enhance the quality of its own AI researchers, toward offering solutions to society.

Overview of RIKEN AIP-Toshiba Collaboration Centre (RATCC)

1. Location:      Nihonbashi 1-chome Mitsui Building, 15th floor, 1-4-1 Nihonbashi,
                        Chuo-ku, Tokyo 103-0027, Japan

2. Term:           April 1, 2017 to March 31, 2022

3. Toshiba’s Main Themes for the R&D of AI
(1) Improving Plant Productivity
As the scale and complexity of factory and power plant systems grow, it is becoming increasingly difficult to pinpoint causes of malfunctions and breakdowns—to a point beyond current management capacity. Toshiba sees the solution in technology that automatically analyzes and optimizes the systems based on big data analysis.

(2) Improving Intellectual Productivity
Companies must ensure continuity of knowledge and institutional memory to maximize the usefulness of their resources. This is becoming increasingly difficult as the knowledge base widens and older, highly capable professionals retire. Toshiba recognizes that technology for mining and excavating know-how from vast amounts of past data, and even discovering previously unrecognized results, provides a solution for building a repository of knowledge.

(3) Automated Mobility/Robotics
While AI technology can now control moving objects, it still faces problems in dealing with unexpected situations and environments, such as unexpected obstacles and sudden wind and rain. Toshiba envisions autonomous decision-making technology that is able to operate in different environments.

Advanced Integrated Intelligence Platform
To utilize enormous data from documents, e-mail, social media as “intelligence”, by extracting necessary information which people want or need, and utilize it to increase the value of applications, and services and customer satisfaction.