SIEVE: Structure-Aware Data Selection for Imitation Learning with VLA Models

Hugging Face Daily Papers Papers
data-selection imitation-learning vision-language-action robot-learning visuo-motor-primitives

Summary

SIEVE is a structure-aware data selection method for vision-language-action imitation learning that identifies reusable visuo-motor primitives and transition interfaces, outperforming full-data training with only 50% of demonstrations and training steps.

Vision-Language-Action (VLA) models are typically trained by imitation learning on large-scale robot demonstration datasets, but more data does not necessarily yield better policies due to redundancy, noise, and uneven coverage. Existing data selection methods often assess demonstrations at either the trajectory or state-action level, missing the reusable structures that compose long-horizon behaviors. In this paper, we propose SIEVE, a structure-aware data selection method for VLA imitation learning. SIEVE views demonstrations as compositions of reusable primitives and transition interfaces. It first discovers visuo-motor primitives from segmented trajectories, then allocates selection budgets to composition patterns by maximizing reuse-aware structural exposure under diminishing returns. Finally, it selects medoid trajectories within each composition-pattern bucket to retain central, stable, and imitation-friendly demonstrations. Experiments across multiple datasets, benchmarks, and VLA models show that SIEVE consistently outperforms competitive data selection baselines. Notably, SIEVE can surpass full-data training while using only 50% of demonstrations and 50% of training steps, suggesting that reusable structure, captured through primitives and transitions, is an important signal for efficient VLA imitation learning.
Original Article
View Cached Full Text

Cached at: 07/08/26, 02:47 AM

Paper page - SIEVE: Structure-Aware Data Selection for Imitation Learning with VLA Models

Source: https://huggingface.co/papers/2607.06442

Abstract

SIEVE is a structure-aware data selection method for vision-language-action imitation learning that identifies reusable visuo-motor primitives and transition interfaces to improve policy learning efficiency.

Vision-Language-Action (VLA) models are typically trained byimitation learningon large-scale robot demonstration datasets, but more data does not necessarily yield better policies due to redundancy, noise, and uneven coverage. Existingdata selectionmethods often assess demonstrations at either the trajectory or state-action level, missing the reusable structures that compose long-horizon behaviors. In this paper, we propose SIEVE, a structure-awaredata selectionmethod for VLAimitation learning. SIEVE views demonstrations as compositions of reusable primitives andtransition interfaces. It first discoversvisuo-motor primitivesfrom segmented trajectories, then allocates selection budgets tocomposition patternsby maximizing reuse-awarestructural exposureunderdiminishing returns. Finally, it selectsmedoid trajectorieswithin each composition-pattern bucket to retain central, stable, and imitation-friendly demonstrations. Experiments across multiple datasets, benchmarks, and VLA models show that SIEVE consistently outperforms competitivedata selectionbaselines. Notably, SIEVE can surpass full-data training while using only 50% of demonstrations and 50% of training steps, suggesting that reusable structure, captured through primitives and transitions, is an important signal for efficient VLAimitation learning.

View arXiv pageView PDFGitHubAdd to collection

Get this paper in your agent:

hf papers read 2607\.06442

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2607.06442 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2607.06442 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2607.06442 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

InSight: Self-Guided Skill Acquisition via Steerable VLAs

Hugging Face Daily Papers

InSight presents a framework for autonomous skill acquisition in vision-language-action (VLA) models by enabling steerability at the primitive-action level and using a VLM-guided data flywheel to generate demonstrations, achieving manipulation tasks like block flipping and pouring without human demonstrations.

IntentVLA: Short-Horizon Intent Modeling for Aliased Robot Manipulation

Hugging Face Daily Papers

IntentVLA is a history-conditioned visual-language-action framework that improves robot imitation learning stability by encoding short-horizon intents from visual observations, addressing challenges from partial observability and ambiguous observations. It also introduces AliasBench, an ambiguity-aware benchmark for evaluating such methods.

Hierarchical Advantage Weighting for Online RL Fine-Tuning of VLAs from Sparse Episode Outcomes

Hugging Face Daily Papers

This paper proposes Hierarchical Advantage-Weighted Behavior Cloning (HABC) for fine-tuning Vision-Language-Action (VLA) policies using online reinforcement learning with sparse binary episode outcomes. HABC separates viability and efficiency objectives via adaptive critic heads and intervention-aware credit assignment, significantly improving success rates on contact-rich bimanual manipulation tasks.