Social Hierarchical Learning

Abstract—I present the outline for my dissertation work, Social Hierarchical Learning (SHL). SHL leverages and extends the capabilities of state-of-the-art hierarchical learning (HL) systems to operate under realistic human-robot interaction domains. SHL is designed for multi-agent coordination with humans in the loop. Traditional HL systems excel when presented tasks in which the agent has full environmental awareness and control, yet such systems are not intended to handle elements outside the agent’s control interacting within its workspace. SHL is designed to provide the required flexibility in task decomposition and assignment for successful human-robot collaboration.

• Type: Journal
• Year: 2015
• Where: AAAI
• Authors: Bradley Hayes
• Keywords: AAAI 2015, Intelligent Agents