Composable Action-Conditioned Predictors: Flexible Off-Policy Learning for Robot Navigation
This addresses the challenge of flexible, multi-task learning for general-purpose robots, though it appears incremental by building on existing off-policy and predictive modeling techniques.
The paper tackles the problem of enabling robots to learn multiple tasks autonomously without predefined rewards by proposing a framework that learns event cues from off-policy data and combines them flexibly at test time, demonstrating that simulated and real-world robotic cars can train without human labels and accomplish various tasks.
A general-purpose intelligent robot must be able to learn autonomously and be able to accomplish multiple tasks in order to be deployed in the real world. However, standard reinforcement learning approaches learn separate task-specific policies and assume the reward function for each task is known a priori. We propose a framework that learns event cues from off-policy data, and can flexibly combine these event cues at test time to accomplish different tasks. These event cue labels are not assumed to be known a priori, but are instead labeled using learned models, such as computer vision detectors, and then `backed up' in time using an action-conditioned predictive model. We show that a simulated robotic car and a real-world RC car can gather data and train fully autonomously without any human-provided labels beyond those needed to train the detectors, and then at test-time be able to accomplish a variety of different tasks. Videos of the experiments and code can be found at https://github.com/gkahn13/CAPs