Learning a Thousand Tasks in a Day

TL;DR

This paper introduces a decomposition and retrieval-based imitation learning method called MT3, enabling robots to learn 1,000 tasks efficiently from minimal demonstrations, significantly surpassing traditional methods in data efficiency and generalization.

Contribution

The paper presents a novel decomposition and retrieval approach for imitation learning, achieving high data efficiency and scalability for robot manipulation tasks from very few demonstrations.

Findings

Decomposition improves data efficiency by an order of magnitude in few-demonstration regimes.

Retrieval-based methods outperform behavioral cloning in alignment and interaction phases.

MT3 can learn 1,000 tasks in under 24 hours of human demonstration time.

Abstract

Humans are remarkably efficient at learning tasks from demonstrations, but today's imitation learning methods for robot manipulation often require hundreds or thousands of demonstrations per task. We investigate two fundamental priors for improving learning efficiency: decomposing manipulation trajectories into sequential alignment and interaction phases, and retrieval-based generalisation. Through 3,450 real-world rollouts, we systematically study this decomposition. We compare different design choices for the alignment and interaction phases, and examine generalisation and scaling trends relative to today's dominant paradigm of behavioural cloning with a single-phase monolithic policy. In the few-demonstrations-per-task regime (<10 demonstrations), decomposition achieves an order of magnitude improvement in data efficiency over single-phase learning, with retrieval consistently outperforming behavioural cloning for both alignment and interaction. Building on these insights, we develop Multi-Task Trajectory Transfer (MT3), an imitation learning method based on decomposition and retrieval. MT3 learns everyday manipulation tasks from as little as a single demonstration each, whilst also generalising to novel object instances. This efficiency enables us to teach a robot 1,000 distinct everyday tasks in under 24 hours of human demonstrator time. Through 2,200 additional real-world rollouts, we reveal MT3's capabilities and limitations across different task families. Videos of our experiments can be found on at https://www.robot-learning.uk/learning-1000-tasks.