Learning from the Best: Smoothness-Driven Metrics for Data Quality in Imitation Learning

TL;DR

This paper introduces RINSE, a lightweight, smoothness-based demonstration scoring framework for improving imitation learning data quality, leading to significant success rate improvements on benchmarks and real-world tasks.

Contribution

RINSE provides a policy-architecture-agnostic, smoothness-based demonstration scoring method that enhances data selection and reweighting in imitation learning without costly policy training.

Findings

SAL filtering yields 16% higher success with one-sixth data on RoboMimic.

TED filtering achieves 20% improvement with half the data in real-world manipulation.

RINSE re-ranking improves success by 5.6% on LIBERO-10.

Abstract

In behavioral cloning (BC), policy performance is fundamentally limited by demonstration data quality. Real-world datasets contain trajectories of varying quality due to operator skill differences, teleoperation artifacts, and procedural inconsistencies, yet standard BC treats all demonstrations equally. Existing curation methods require costly policy training in the loop or manual annotation, limiting scalability. We propose RINSE (Ranking and INdexing Smooth Examples), a lightweight framework for scoring demonstrations based on trajectory smoothness that is policy-architecture-agnostic and operates on trajectory data alone, with TED additionally using a phase-boundary/contact signal. Grounded in motor control theory, which establishes smoothness as a hallmark of skilled movement, RINSE uses two complementary metrics: Spectral Arc Length (SAL), a spectral measure of frequency-domain regularity, and Trajectory-Envelope Distance (TED), a spatial measure of contact-aware geometric deviation. We show that smoothness filtering can reduce the conditional action variance of the retained data distribution, with downstream effects that can be amplified by action chunking and compounding error. On RoboMimic benchmarks, SAL filtering achieves 16% higher success using one-sixth of the data. On real-world manipulation, TED filtering achieves 20% improvement with half the data. As a retrieval-stage filter within STRAP on LIBERO-10, RINSE re-ranking improves mean success by 5.6%. As soft weights in Re-Mix domain reweighting, RINSE scores produce domain allocations highly correlated with the learned Re-Mix allocations (Spearman $\rho \geq 0.89$). These results support smoothness as a useful quality signal across filtering, retrieval, and reweighting settings, especially in noisy or heterogeneous data regimes.