Demonstration Sidetracks: Categorizing Systematic Non-Optimality in Human Demonstrations

TL;DR

This paper identifies and categorizes systematic non-optimal behaviors, called demonstration sidetracks, in human demonstrations for robot learning, highlighting their patterns and implications for improving Learning from Demonstration methods.

Contribution

It introduces a novel taxonomy of demonstration sidetracks, providing detailed analysis of their types, distribution, and dependence on control interfaces in human demonstrations.

Findings

Sidetracks are frequent and context-dependent.

Four types of sidetracks identified: Exploration, Mistake, Alignment, Pause.

Control patterns vary with interface and influence demonstration quality.

Abstract

Learning from Demonstration (LfD) is a popular approach for robots to acquire new skills, but most LfD methods suffer from imperfections in human demonstrations. Prior work typically treats these suboptimalities as random noise. In this paper we study non-optimal behaviors in non-expert demonstrations and show that they are systematic, forming what we call demonstration sidetracks. Using a public space study with 40 participants performing a long-horizon robot task, we recreated the setup in simulation and annotated all demonstrations. We identify four types of sidetracks (Exploration, Mistake, Alignment, Pause) and one control pattern (one-dimension control). Sidetracks appear frequently across participants, and their temporal and spatial distribution is tied to task context. We also find that users' control patterns depend on the control interface. These insights point to the need for better models of suboptimal demonstrations to improve LfD algorithms and bridge the gap between lab training and real-world deployment. All demonstrations, infrastructure, and annotations are available at https://github.com/AABL-Lab/Human-Demonstration-Sidetracks.