Visually Robust Adversarial Imitation Learning from Videos with Contrastive Learning

TL;DR

C-LAIfO is a new imitation learning algorithm that uses contrastive learning to create a robust latent space from videos, enabling effective imitation despite visual mismatches between agent and expert videos.

Contribution

The paper introduces C-LAIfO, a contrastive learning-based method for robust imitation learning from videos with visual discrepancies, and demonstrates its effectiveness on robotic tasks.

Findings

Improved imitation performance over baselines

Effective handling of visual mismatches

Successful integration with sparse reward signals

Abstract

We propose C-LAIfO, a computationally efficient algorithm designed for imitation learning from videos in the presence of visual mismatch between agent and expert domains. We analyze the problem of imitation from expert videos with visual discrepancies, and introduce a solution for robust latent space estimation using contrastive learning and data augmentation. Provided a visually robust latent space, our algorithm performs imitation entirely within this space using off-policy adversarial imitation learning. We conduct a thorough ablation study to justify our design and test C-LAIfO on high-dimensional continuous robotic tasks. Additionally, we demonstrate how C-LAIfO can be combined with other reward signals to facilitate learning on a set of challenging hand manipulation tasks with sparse rewards. Our experiments show improved performance compared to baseline methods, highlighting the effectiveness of C-LAIfO. To ensure reproducibility, we open source our code.