Hokkaido University Research Profiles

Japanese
Information and Communication

Multimedia Artificial Intelligence Technology Reaching Social Implementation

Approaching the practical application of AI technology through industry-university collaborative research!

With this research, we are developing artificial intelligence technology for multimedia data, mainly images, video, music, and audio. We are handling data related to medical images, social infrastructure data, materials science and other fields, mainly through industry-university collaborative research.

Content of research

We are not only conducting the world's most advanced artificial intelligence research, but also promoting research in interdisciplinary areas and taking on the challenge of solving real-world problems. Specifically, in medical imaging research, we have collaborated with many medical institutions in Japan to build AI technology that surpasses human diagnostic accuracy. In medical and civil engineering research, we have built Explainable AI (XAI), which not only enables learning of small amounts of data, a challenge in AI research, but also enables explanations of judgment results, making the technology usable in the real world. In recent years, we have also developed human-centric AI technology that can make decisions like humans by introducing information strongly related to human interests, such as human brain activity and eye gaze data, into the AI learning process.

  • State-of-the-art AI research groups are built through this research

  • Attempts for interdisciplinary collaboration and social implementation accelerated by this research

Potential for social implementation

  • ・Multimodal analysis AI technology that does not only analyze multimedia data but also a wide variety of other data.
  • ・AI technology that enables learning from limited data
  • ・Explainable AI (XAI) technology to reveal the basis for decisions
  • ・Knowledge acquisition technology for AI that can inherit the tacit knowledge of skilled workers
  • ・Multimedia search and recommendation using AI technology

Appealing points to industry and local governments

We are currently promoting a variety of industry-academia collaborative research projects to build AI that can be implemented in society and solve real-world problems. In addition to the field of information and communication, we are also very active in industry-university collaborative research with companies in the fields of medicine, social infrastructure maintenance and management, and materials science. We are also conducting joint research with local governments, and have ten years’ experience in demonstration experiments.

Intellectual property related to this research

PCT/JP2008/069193  「類似画像検索装置」
(日本:特許第5322111 号  米国:特許第8,180,162 号  欧州:08847718.7)
本特許に関する基礎理論については、電子情報通信学会論文賞を受賞しています。
2022/5/27Released