[Seminar] by Dr. Katsunari Shibata, Oita University
Dr. Katsunari Shibata, Associate Professor, Oita University
Emergence of Intelligence through End-to-End Reinforcement Learning and Chaos-based Reinforcement Learning towards the Emergence of "Thinking"
On the ground of my 20-year relevant researches, I will propound the significance of "End-to-End Reinforcement Learning" using a (recurrent) neural network from the viewpoint of emergence of intelligence. Towards the ambitious goal of emergence of ”Thinking”, I will also introduce my new fundamental idea that ”explorations” and ”actions” should be unified and the internal ”exploration” grows into ”thinking” through learning. New reinforcement learning I have proposed to realize that uses a chaotic neural network and jumps out of the conventional stochastic action selection approach.