Humanoid Cognitive Robots That Learn by Imitating: Implications for Consciousness Studies
Frontiers in Robotics and AI January 26, 2018 James A. Reggia, Garrett E. Katz, Gregory P. Davis 41 citations
Creating a conscious machine remains controversial and challenging. This work describes a humanoid cognitive robot that learns tasks by imitating human demonstrations, using cause-effect reasoning to infer a demonstrator's intentions rather than merely copying actions. Its cognitive components center on top-down control of working memory, which retains explanatory interpretations constructed during learning. Ongoing work aims to convert this imitation learning system into purely neurocomputational form, including low-level neuromotor components, working memory, and causal reasoning. Based on initial results, top-down cognitive control of working memory—especially its gating mechanisms—is argued to be an important potential computational correlate of consciousness in humanoid robots. Developing such neurocognitive control systems provides a credible route to ultimately developing a phenomenally conscious machine.