Learning Quickly to Plan Quickly Using Modular Meta-Learning

Multi-object manipulation problems in continuous state and action spaces can be solved by planners that search over sampled values for the continuous parameters of operators. The efficiency of these planners depends critically on the effectiveness of the samplers used, but effective sampling in turn depends on details of the robot, environment, and task. Our strategy is to learn functions called speciatizers that generate values for continuous operator parameters, given a state description and values for the discrete parameters. Rather than trying to learn a single specializer for each operator from large amounts of data on a single task, we take a modular meta-learning approach. We train on multiple tasks and learn a variety of specializers that, on a new task, can be quickly adapted using relatively little data - thus, our system learns quickly to plan quickly using these specializers. We validate our approach experimentally in simulated 3D pick-and-place tasks with continuous state and action spaces. Visit http://tinyurl.com/chitnis-icra-19 for a supplementary video.

Authors

• Rohan Chitnis
• Leslie Pack Kaelbling
• Tomás Lozano-Pérez

Keywords

• Task analysis
• Planning
• Robots
• Skeleton
• Search problems
• Companies
• Neural networks
• Large Amount Of Data
• Continuous Parameters
• Continuous Action Space
• Continuous State Space
• Neural Network
• Random Sampling
• Global Status
• Continuous Values
• Path Planning
• Training Tasks
• Problem Instances
• Complete Solution
• Subset Selection
• Planning Problem
• Final Task
• Search Effort
• Object Pose
• Object Geometry
• Action Instances
• End-effector Pose
• Robot Configuration
• Discrete Objects
• Deployment Time
• Rating Task
• Functional Form
• Aspects Of Planning