Effective Robot Task Learning by focusing on Task-relevant objects

In a Robot Learning from Demonstration framework involving environments with many objects, one of the key problems is to decide which objects are relevant to a given task. In this paper, we analyze this problem and propose a biologically-inspired computational model that enables the robot to focus on the task-relevant objects. To filter out incompatible task models, we compute a Task Relevance Value (TRV) for each object, which shows a human demonstrator's implicit indication of the relevance to the task. By combining an intentional action representation with ‘motionese’ [2], our model exhibits recognition capabilities compatible with the way that humans demonstrate. We evaluate the system on demonstrations from five different human subjects, showing its ability to correctly focus on the appropriate objects in these demonstrations.

Authors

• Kyuhwa Lee
• Jinhan Lee
• Andrea Thomaz
• Aaron F. Bobick

Keywords

• Humans
• Intelligent robots
• Computational modeling
• Educational robots
• Learning systems
• Computer architecture
• Object oriented modeling
• USA Councils
• Machine learning
• Biology computing
• Learning Task
• Task-relevant Objects
• Relevant Tasks
• Inverse Reinforcement Learning
• Robot Learning
• Left Side
• System State
• Human Behavior
• Sequence Of Actions
• Visual Attention
• Forward Model
• Inverse Model
• Error Signal
• Confidence Value
• Part Of The Task
• Objects In The Scene
• Human Users
• Current Time Step
• Human Education
• Beginning Of The Task
• Task Representations