IIFL: Implicit Interactive Fleet Learning from Heterogeneous Human Supervisors

TL;DR

IIFL introduces an advanced imitation learning algorithm that effectively learns from diverse human supervisors, significantly improving success rates and reducing human effort in robotic tasks through implicit modeling and uncertainty quantification.

Contribution

The paper presents IIFL, a novel implicit interactive fleet learning algorithm that leverages energy-based models and Jeffreys divergence for uncertainty, enhancing learning from heterogeneous human demonstrations.

Findings

Achieves 2.8x higher success rate in simulations.

Attains 4.5x higher human effort efficiency in physical tasks.

Outperforms existing methods like IFL and IBC.

Abstract

Imitation learning has been applied to a range of robotic tasks, but can struggle when robots encounter edge cases that are not represented in the training data (i.e., distribution shift). Interactive fleet learning (IFL) mitigates distribution shift by allowing robots to access remote human supervisors during task execution and learn from them over time, but different supervisors may demonstrate the task in different ways. Recent work proposes Implicit Behavior Cloning (IBC), which is able to represent multimodal demonstrations using energy-based models (EBMs). In this work, we propose Implicit Interactive Fleet Learning (IIFL), an algorithm that builds on IBC for interactive imitation learning from multiple heterogeneous human supervisors. A key insight in IIFL is a novel approach for uncertainty quantification in EBMs using Jeffreys divergence. While IIFL is more computationally expensive than explicit methods, results suggest that IIFL achieves a 2.8x higher success rate in simulation experiments and a 4.5x higher return on human effort in a physical block pushing task over (Explicit) IFL, IBC, and other baselines.