Design of Mimesis Loop by Ergodic Hidden Markov Model

摘要

Relationship between motion recognition and generation has been expected to be indispensible for acquisition of intelligence. Mimesis Loop Model for implementing the intelligence for abstracting motion patterns into proto symbols and generating motions from these symbols has been proposed. In this paper we introduce ergodic hidden markov model in order to configure a large scale models which can recognize and generate various motion patterns.