MinInter: Minimizing Trajectory Interpolation During Data Augmentation for Imitation Learning

2026-06-22

AI Summary

This work addresses the degradation of synthetic data quality in imitation learning caused by non-expert interpolation segments introduced through trajectory-level data augmentation. To mitigate this issue, the authors propose MinInter, a novel method that explicitly minimizes interpolation as a trajectory selection criterion: for each initial state, it selects the expert trajectory requiring the least interpolation for recombination, thereby significantly reducing non-expert transitions. This strategy seamlessly integrates with existing data generation frameworks and demonstrates consistent improvements across 12 tasks and 26 variants in the MimicGen benchmark, markedly enhancing both synthetic data fidelity and policy success rates—particularly in contact-intensive, long-horizon, and high-variance scenarios.

Abstract

Imitation learning enables robots to acquire complex manipulation skills from demonstrations, but its effectiveness is limited by the cost of collecting high-quality data. Trajectory-level data augmentation methods alleviate this challenge by recombining expert demonstrations under varied initial states. However, such methods typically insert interpolations or other non-expert transition segments between disjoint parts, and such non-expert segments could reduce the quality of the generated data. This paper introduces Minimizing Interpolation (MinInter), an effective trajectory selection method that, for each sampled initial configuration, chooses the source demonstration requiring the least interpolation to form a complete trajectory. By explicitly minimizing interpolations during data generation, MinInter produces higher-quality synthetic demonstrations while remaining compatible with existing data generation frameworks. Experiments on 12 manipulation tasks with 26 variants from the MimicGen benchmark show that MinInter consistently improves both data generation success rates and policy success rates, with the largest gains on contact-rich, long-horizon and high-variance settings. Compared to the recent SkillGen framework, MinInter achieves higher policy success rates despite its conceptual simplicity, underscoring the value of interpolation minimization for data augmentation.

Problem

• Imitation Learning
• Data Augmentation
• Trajectory Interpolation
• Demonstration Quality
• Robot Manipulation

Innovation

• trajectory interpolation minimization
• imitation learning
• data augmentation
• synthetic demonstration generation
• robot manipulation