TFC×TEL Collaboration Event: Special Lecture by 2025 Kyoto Prize Laureate
Society of Instrument and Control Engineers (SICE), Systems and Information Division (SANDI), SSI2025, Special lecture
Information
Shun–ichi Amari has conducted pioneering research in artificial neural networks and established the field of information geometry, which studies statistical models using the techniques of differential geometry, thereby proposing many important theories. His contributions to both theory and application, influencing various fields, are of major significance. With this citation, Professor Shun–ichi Amari is awarded Kyoto Prize on November 10, 2025.
The SSI2025 of the Society of Instrument and Control Engineers is held at Tohoku University's Aobayama Campus from 12 to 14 November 2025. The special lecture by Professor Shunichi Amari at SSI2025 is co–hosted by the Tohoku Forum for Creativity, Tohoku University and made available to members of Tohoku University. We extend our gratitude to the Systems and Information Division of the Society of Instrument and Control Engineers for their kind cooperation.
Date
Friday, November 14, 2025, 13:15–14:45
Venue
Main Conference Room, Centersquare CENTER HALL, Aobayama Campus, Tohoku University [Access]
*Broadcasted to Main Lecture Hall, or Online
Capacity
100 in the Main Conference Room
250 in the Main Lecture Hall
200 online
Speaker
Shun–ichi Amari (Specially Appointed Professor, Teikyo University / Honorary Science Advisor, RIKEN)
Title of talk
History and current state of artificial intelligence, and an outlook of future civilisation
Abstract
- The development of AI has been astonishing. Moreover, it continues to advance rapidly, poised to exert significant influence not only on academia but also on society and civilisation. This lecture provides a survey on the history of AI comparing with the human brain. Furthermore, it provides an overview of its current state.
A major issue is the lag in theoretical understanding. Theories addressing scaling laws that become apparent as scale increases, or phenomena like emergence, remain underdeveloped. I speculate this might relate to singularities in the parameter space. In deep neural networks, the increase of the number of parameters makes the mesh of singularities cover densely the parameter space. Optimisation problems in such spaces have been scarcely discussed to date. Singular structures also appear in control theory in the space of rational transition functions.
AI is immensely useful, and we cannot afford not to utilise it. On the other hand, it may strip away human cognitive abilities and profoundly impact civilisation. Considering the dangers of AI and addressing them is an extremely important challenge.
Registration
Necessary and permitted only to the members of Tohoku University (possessing the mail addresses with tohoku.ac.jp).
Registration deadline: Friday, November 12, 2025 12:00 → Thursday, November 13, 2025 12:00
*Registration deadline has been extended.
– Closed
Time Schedule
- 13:00
- Door open
- 13:15
- Opening
- 13:15 – 14:45
- Shun–ichi Amari (Specially Appointed Professor, Teikyo University / Honorary Science Advisor, RIKEN)
- History and current state of artificial intelligence, and an outlook of future civilisation
- 14:45
- Closing
Video
Hosted by
Society of Instrument and Control Engineers, Systems and Information Division
Co–hosted by
- Graduate School of Information Sciences, Tohoku University
- Center for Data–driven Science and Artificial Intelligence, Tohoku University
- Society of Instrument and Control Engineers (SICE), Tohoku Chapter
- United Centers for Advanced Research and Translational Medicine, Tohoku University
- Tohoku University Smart Aging Research Center
- Tohoku Forum for Creativity, Tohoku University
Contact
Tohoku Forum for Creativity
tfc_se@grp.tohoku.ac.jp