M. Kikuchi

About

M. Kikuchi explores the intersection of developmental robotics and humanoid cognition, focusing on how machines can acquire language and motor skills through active, curiosity-driven learning. Drawing inspiration from infant development, Kikuchi’s most influential work proposes a lexical acquisition model that mimics how human children learn words from minimal, ambiguous input—achieving robust mapping between meaning and utterance despite numerous possible interpretations. This research, published in 2006 and garnering 8 citations, highlights the role of attention and learning biases in enabling humanoids to build lexicons autonomously. Additionally, Kikuchi’s work on visuo-motor learning (2005, 2 citations) introduces a method for humanoids to generate behaviors by learning sensorimotor maps that link optic flow in their visual field to motion parameters. By leveraging forward and inverse models, these robots can determine appropriate actions without explicit programming. Kikuchi’s contributions advance the field of cognitive robotics, offering a framework for creating more adaptive, curious machines that learn like humans—a foundational step toward truly intelligent humanoids.

Research Focus

Key Achievements

Top Papers

1. 1
   How can humanoid acquire lexicon? -active approach by attention and learning biases based on curiosity-

   8 citations · 2006
2. 2
   Visuo-motor learning for behavior generation of humanoids

   2 citations · 2005

Key Collaborators

Contact & Links