Father of Information Geometry Honored
The Technische Universität of Berlin has awarded an honorary doctorate to the eminent Japanese scientist Prof. Shun-ichi Amari in recognition of his groundbreaking contributions to the mathematical foundations of Artificial Intelligence. Prof. Amari is Professor Emeritus at the University of Tokyo, Japan’s most prestigious university, and an Honorary Fellow at the RIKEN Institute, one of Japan’s leading national research organizations, comparable to the Helmholtz Association in Germany. He is regarded as a world-class scientist who has shaped and inspired modern AI research for decades. The honorary degree was presented to him during a ceremony in Tokyo on November 19th by Klaus-Robert Müller, Professor of Machine Learning at TU Berlin and Co-Director of BIFOLD.
Prof. Amari has made major contributions to the theoretical foundations of machine learning and theoretical neuroscience. His work has been instrumental in enabling the success of today’s artificial intelligence in both science and industry. In 1967, he developed the first learning algorithm for training multilayer neural networks using so called “stochastic gradient descent” and provided a comprehensive mathematical analysis of its properties. Due to limited computing power at the time, this method could not yet be applied in practice. Nearly twenty years later, the approach was independently rediscovered in a similar form as the “backpropagation algorithm”, now a central tool of deep learning. His early research on associative memories was also far ahead of its time.
Prof. Amari is furthermore regarded as the father of information geometry. He applied methods from differential geometry to study statistical models as well as learning and optimization methods. His theory builds a bridge between differential geometry and statistics. The field he founded is now internationally established, with its own conferences and workshops, and has significantly advanced our understanding of machine learning methods, particularly of neural networks.
Beyond his scientific excellence, Prof. Amari is an inspiring teacher and mentor who has shaped generations of researchers, including Prof. Klaus-Robert Müller, who worked with him as a visiting scholar at the University of Tokyo in 1994/1995 on the theory of neural networks. Many of Amari`s former students now hold leading positions in research and industry.
For his outstanding achievements, Prof. Amari was awarded, among other honors, the Order of Culture in 2019, Japan’s highest distinction for contributions to the arts, literature, science, and technology, presented personally by the Emperor of Japan. On November 10th, 2025, he received the prestigious “Kyoto Prize”.
Prof. Amari has maintained a long-standing and highly productive collaboration with TU Berlin, including numerous visits and research stays. With the awarding of the honorary doctorate, TU Berlin honors an exceptional researcher, teacher, and pioneer of Artificial Intelligence whose work has fundamentally transformed our understanding of learning, information, and intelligence.