Unsupervised Feature Learning for Manipulation with Contrastive Domain Randomization

TL;DR

This paper introduces a modified contrastive learning approach with domain randomization for robotic manipulation, resulting in more robust visual features that transfer better from simulation to real-world scenarios.

Contribution

It proposes a simple modification to contrastive loss that improves invariance to irrelevant visual properties during unsupervised feature learning.

Findings

Features are more robust to visual domain variations.

Improved transferability of learned features to real-world tasks.

Effective for both rigid and non-rigid objects.

Abstract

Robotic tasks such as manipulation with visual inputs require image features that capture the physical properties of the scene, e.g., the position and configuration of objects. Recently, it has been suggested to learn such features in an unsupervised manner from simulated, self-supervised, robot interaction; the idea being that high-level physical properties are well captured by modern physical simulators, and their representation from visual inputs may transfer well to the real world. In particular, learning methods based on noise contrastive estimation have shown promising results. To robustify the simulation-to-real transfer, domain randomization (DR) was suggested for learning features that are invariant to irrelevant visual properties such as textures or lighting. In this work, however, we show that a naive application of DR to unsupervised learning based on contrastive estimation does not promote invariance, as the loss function maximizes mutual information between the features and both the relevant and irrelevant visual properties. We propose a simple modification of the contrastive loss to fix this, exploiting the fact that we can control the simulated randomization of visual properties. Our approach learns physical features that are significantly more robust to visual domain variation, as we demonstrate using both rigid and non-rigid objects.