Acquisition of Language and Action Concepts from Multimodal Information

Mon, 01 Jan 2024 00:00:00 +0000

Through an integrated cognitive architecture that fuses multimodal information (visual, auditory, and bodily), we study how robots simultaneously acquire and ground action and language concepts (symbol grounding). Based on probabilistic generative models, we build frameworks in which concepts and words emerge jointly from embodied experience.

Curiosity-Inspired Complementary Learning of Actions and Concepts

Mon, 01 Jan 2024 00:00:00 +0000

Inspired by curiosity-driven exploration in infants, we build learning models in which robot action acquisition and concept formation progress in a mutually reinforcing way. Through computational models and robot experiments, we investigate how early motor reflexes such as grasping interact with the physical properties of the environment to shape both action repertoires and concept representations.

Concept-Learning | Takato Horii | Osaka University

Acquisition of Language and Action Concepts from Multimodal Information

Curiosity-Inspired Complementary Learning of Actions and Concepts