[2M6-OS-19d] 世界モデルと知能(4/4)

[2M6-OS-19d-04] A Whole Brain Probabilistic Generative Model

Approach for Cognitive Architecture of Developmental Robots

〇Tadahiro Taniguchi Taniguchi¹, Hiroshi Yamakawa^2,6, Takayuki Nagai³, Kenji Doya⁴, Masamichi Sakagami⁵, Masahiro Suzuki⁶, Tomoaki Nakamura⁷, Akira Taniguchi¹ (1. Ritsumeikan University, 2. The Whole Brain Architecture Initiative, 3. Osaka University, 4. OIST, 5. Tamagawa University, 6. The University of Tokyo, 7. The University of Electro-Communications)

Keywords:Cognitive architecture, Probabilistic generative model, Brain-inspired artificial intelligence, Cognitive developmental robotics, Symbol emergence in robotics

This paper describes a concept for building a human-like integrative artificial cognitive system, that is, an artificial general intelligence (AGI). The aim is to develop a computational model that enables an artificial system to achieve cognitive development
by referring human brain structure. The approach is called a whole brain probabilistic generative model (WB-PGM). WB-PGM is based on two ideas: (1) brain-inspired AI, learning human brain architecture to build human-level intelligence, and (2) a probabilistic generative model(PGM)-based cognitive architecture to develop a cognitive system for developmental robots by integrating PGMs.