Embedded Files
By using our Multi Agent Reinforcement Learning, we will realize the real autonomous Artificial Intelligence.
What is Multi Agent Reinforcement Learning?
What is Multi Agent Reinforcement Learning?
Conventional Reinforcement Learning that is the way to realize the AI with autonomous trial and error function for learning, can't interact with other agents because of the limitation of modeling for the environment.
Multi Agent Reinforcement Learning is the way to attack that problem by introducing the knowledge of Game Theory that is one of an important domain of Economics, and it is regarded as an important element for autonomous Artificial Intelligence, however this domain is still in the long way to be developed.
By using our original algorithm of Multi Agent Reinforcement Learning, we will realize the real autonomous Artificial Intelligence.
POSG
POSG
Technically saying, our modeling adopts "Partially Observable Stochastic Games" assumption. We are analyzing the theoretical property of this model without spoiling the generality of assumptions.
Non-parametric Bayes
Non-parametric Bayes
Non-parametric Bayes is also a hot technique of Statistics / Machine Learning domain. We are adopting Non-parametric Bayes as elements of our Multi Agent Reinforcement Learning for both enhancing the precision and keeping the generality of assumptions.
KIMURA Yu
KIMURA Yu
Ba(Econ), MBA, and candidate of PhD(Info) at Kyoto University.
By combining the knowledge of Economics to the Artificial Intelligence domain, researching the brand-new original algorithm.
Inquiry
Inquiry
To know about our project more, please contact with:
info[at]asia.cauchye.com
Page updated
Report abuse