JSAI Technical Report, Type 2 SIG
Online ISSN : 2436-5556
The reward made with pleasant and unpleasant will reach general purpose
Yuki KATO
Author information
RESEARCH REPORT / TECHNICAL REPORT FREE ACCESS

2019 Volume 2019 Issue AGI-011 Pages 03-

Details
Abstract

In recent years, the development of AI field is remarkable. In particular, it is remembrance that Alpha Go beat the world champion of Go. In 2019, the practical use of AI in the business area is advanced and what AI can solve or not will be clearly with people who are not specialists. Metastasis learning is one of the unresolved issues and this issue leads to the acquisition of versatility. Due to Metastasis learning, introduction of pleasant and unpleasant to AI is effective. In this paper, pleasant has nature of "dissolution of unpleasant", and therefore pleasant is based on unpleasant. It brings distinction to reward by unpleasant and ability to distinguish brings AI to unique purpose for each task.

Content from these authors
© 2019 Authors
Previous article Next article
feedback
Top