MIRROR: Differentiable Deep Social Projection for Assistive Human-Robot Communication

TL;DR

MIRROR is a differentiable social projection method that rapidly learns human models from demonstrations and enhances assistive communication in complex, high-dimensional environments, improving safety and robustness.

Contribution

It introduces MIRROR, a novel approach leveraging self-models learned via reinforcement learning for quick human modeling and effective assistive communication in shared-control settings.

Findings

Faster learning of human models compared to existing methods.

Effective assistive communication improves driving safety.

Scales to complex, high-dimensional environments.

Abstract

Communication is a hallmark of intelligence. In this work, we present MIRROR, an approach to (i) quickly learn human models from human demonstrations, and (ii) use the models for subsequent communication planning in assistive shared-control settings. MIRROR is inspired by social projection theory, which hypothesizes that humans use self-models to understand others. Likewise, MIRROR leverages self-models learned using reinforcement learning to bootstrap human modeling. Experiments with simulated humans show that this approach leads to rapid learning and more robust models compared to existing behavioral cloning and state-of-the-art imitation learning methods. We also present a human-subject study using the CARLA simulator which shows that (i) MIRROR is able to scale to complex domains with high-dimensional observations and complicated world physics and (ii) provides effective assistive communication that enabled participants to drive more safely in adverse weather conditions.