Message from the project leader
Artificial intelligence and brain science have had a swinging relationship of convergence and divergence. Views like “To create intelligent machines, we should not be constrained by biological implementation.” and “As there already is a superb realization by the brain, why not learn from its mechanisms.” both sound reasonable. Recently, as artificial intelligence shifted from copying human knowledge and skills to learning from big data, the remarkable success of deep neural networks has re-evoked wide interests in brain-like artificial intelligence.
Also on the side of brain science, with the advances in sequencing, imaging, and other high-throughput technologies, the use of statistical learning and artificial intelligence is becoming a must. The aim of this research project is to promote dialogues and collaborations among researchers in artificial intelligence and brain science and to facilitate evolution of novel research.
Learning in deep neural networks has been the best example of brain-inspired artificial intelligence, but we postulate that there are many other domains where we can learn from the brain. We have organized this new project by recruiting active members from both camps.
In a complex distributed system like the brain, how can appropriate modules be activated and combined to address varieties of problems it encounters? The principle of self-organization at the whole-brain scale is a big open question in brain science. Its understanding would allow us to build artificial general intelligence that can flexibly adapt to varieties of situations and problems. Comparison of such advanced artificial intelligence and humans would bring us deeper understanding of the nature of the brain.
Professor, Neural Computation Unit
Okinawa Institute of Science and Technology Graduate University