Third-Person Visual Imitation Learning via Decoupled Hierarchical Controller

TL;DR

This paper introduces a hierarchical third-person imitation learning framework enabling robots to learn manipulation tasks from a single human demonstration video, by decoupling task understanding from control execution.

Contribution

It proposes a hierarchical model with a high-level goal generator and a low-level controller, explicitly decoupling task intent from control, for learning from third-person videos in unseen scenarios.

Findings

Successfully applied to real robot manipulation tasks

Achieved generalization to unseen objects and scenarios

Operated using only raw image observations

Abstract

We study a generalized setup for learning from demonstration to build an agent that can manipulate novel objects in unseen scenarios by looking at only a single video of human demonstration from a third-person perspective. To accomplish this goal, our agent should not only learn to understand the intent of the demonstrated third-person video in its context but also perform the intended task in its environment configuration. Our central insight is to enforce this structure explicitly during learning by decoupling what to achieve (intended task) from how to perform it (controller). We propose a hierarchical setup where a high-level module learns to generate a series of first-person sub-goals conditioned on the third-person video demonstration, and a low-level controller predicts the actions to achieve those sub-goals. Our agent acts from raw image observations without any access to the full state information. We show results on a real robotic platform using Baxter for the manipulation tasks of pouring and placing objects in a box. Project video and code are at https://pathak22.github.io/hierarchical-imitation/