Generalizing to New Tasks via One-Shot Compositional Subgoals

by   Xihan Bian, et al.

The ability to generalize to previously unseen tasks with little to no supervision is a key challenge in modern machine learning research. It is also a cornerstone of a future "General AI". Any artificially intelligent agent deployed in a real world application, must adapt on the fly to unknown environments. Researchers often rely on reinforcement and imitation learning to provide online adaptation to new tasks, through trial and error learning. However, this can be challenging for complex tasks which require many timesteps or large numbers of subtasks to complete. These "long horizon" tasks suffer from sample inefficiency and can require extremely long training times before the agent can learn to perform the necessary longterm planning. In this work, we introduce CASE which attempts to address these issues by training an Imitation Learning agent using adaptive "near future" subgoals. These subgoals are recalculated at each step using compositional arithmetic in a learned latent representation space. In addition to improving learning efficiency for standard long-term tasks, this approach also makes it possible to perform one-shot generalization to previously unseen tasks, given only a single reference trajectory for the task in a different environment. Our experiments show that the proposed approach consistently outperforms the previous state-of-the-art compositional Imitation Learning approach by 30


page 1

page 4


Plan Arithmetic: Compositional Plan Vectors for Multi-Task Control

Autonomous agents situated in real-world environments must be able to ma...

Transfering Hierarchical Structure with Dual Meta Imitation Learning

Hierarchical Imitation Learning (HIL) is an effective way for robots to ...

Hierarchical Few-Shot Imitation with Skill Transition Models

A desirable property of autonomous agents is the ability to both solve l...

Long-term Microscopic Traffic Simulation with History-Masked Multi-agent Imitation Learning

A realistic long-term microscopic traffic simulator is necessary for und...

Continuous Relaxation of Symbolic Planner for One-Shot Imitation Learning

We address one-shot imitation learning, where the goal is to execute a p...

SKILL-IL: Disentangling Skill and Knowledge in Multitask Imitation Learning

In this work, we introduce a new perspective for learning transferable c...

Meta Adaptation using Importance Weighted Demonstrations

Imitation learning has gained immense popularity because of its high sam...

Please sign up or login with your details

Forgot password? Click here to reset