Academic Bibliography

 Search 200 years of publications by Ghent University researchers.
Advanced search
2 files | 15.72 MB
Add to list
  1. Search results
  2. Learning self-supervised task progres...

Learning self-supervised task progression metrics : a case of cloth folding

Andreas Verleysen (UGent) , Matthijs Biondina (UGent) and Francis wyffels (UGent)
(2023) APPLIED INTELLIGENCE. 53(2). p.1725-1743
Author
Organization
Project
Abstract
An important challenge for smart manufacturing systems is finding relevant metrics that capture task quality and progression for process monitoring to ensure process reliability and safety. Data-driven process metrics construct features and labels from abundant raw process data, which incurs costs and inaccuracies due to the labelling process. In this work, we circumvent expensive process data labelling by distilling the task intent from video demonstrations. We present a method to express the task intent in the form of a scalar value by aligning a self-supervised learned embedding to a small set of high-quality task demonstrations. We evaluate our method on the challenging case of monitoring the progress of people folding clothing. We demonstrate that our approach effectively learns to represent task progression without manually labelling sub-steps or progress in the videos. Using case-based experiments, we find that our method learns task-relevant features and useful invariances, making it robust to noise, distractors and variations in the task and shirts. The experimental results show that the proposed method can monitor processes in domains where state representation is inherently challenging.
Keywords
Artificial Intelligence, Process monitoring, Deformable object manipulation, Contrastive learning, Learning from demonstrations, Semantic representation

Downloads

  • Download
    main.pdf
    • full text (Accepted manuscript)
    • |
    • open access
    • |
    • PDF
    • |
    • 11.71 MB
  • Download
    (...).pdf
    • full text (Published version)
    • |
    • UGent only
    • |
    • PDF
    • |
    • 4.01 MB

Citation

Please use this url to cite or link to this publication:

MLA
Verleysen, Andreas, et al. “Learning Self-Supervised Task Progression Metrics : A Case of Cloth Folding.” APPLIED INTELLIGENCE, vol. 53, no. 2, 2023, pp. 1725–43, doi:10.1007/s10489-022-03466-8.
APA
Verleysen, A., Biondina, M., & wyffels, F. (2023). Learning self-supervised task progression metrics : a case of cloth folding. APPLIED INTELLIGENCE, 53(2), 1725–1743. https://doi.org/10.1007/s10489-022-03466-8
Chicago author-date
Verleysen, Andreas, Matthijs Biondina, and Francis wyffels. 2023. “Learning Self-Supervised Task Progression Metrics : A Case of Cloth Folding.” APPLIED INTELLIGENCE 53 (2): 1725–43. https://doi.org/10.1007/s10489-022-03466-8.
Chicago author-date (all authors)
Verleysen, Andreas, Matthijs Biondina, and Francis wyffels. 2023. “Learning Self-Supervised Task Progression Metrics : A Case of Cloth Folding.” APPLIED INTELLIGENCE 53 (2): 1725–1743. doi:10.1007/s10489-022-03466-8.
Vancouver
1.
Verleysen A, Biondina M, wyffels F. Learning self-supervised task progression metrics : a case of cloth folding. APPLIED INTELLIGENCE. 2023;53(2):1725–43.
IEEE
[1]
A. Verleysen, M. Biondina, and F. wyffels, “Learning self-supervised task progression metrics : a case of cloth folding,” APPLIED INTELLIGENCE, vol. 53, no. 2, pp. 1725–1743, 2023.
@article{8752294,
  abstract     = {{An important challenge for smart manufacturing systems is finding relevant metrics that capture task quality and progression for process monitoring to ensure process reliability and safety. Data-driven process metrics construct features and labels from abundant raw process data, which incurs costs and inaccuracies due to the labelling process. In this work, we circumvent expensive process data labelling by distilling the task intent from video demonstrations. We present a method to express the task intent in the form of a scalar value by aligning a self-supervised learned embedding to a small set of high-quality task demonstrations. We evaluate our method on the challenging case of monitoring the progress of people folding clothing. We demonstrate that our approach effectively learns to represent task progression without manually labelling sub-steps or progress in the videos. Using case-based experiments, we find that our method learns task-relevant features and useful invariances, making it robust to noise, distractors and variations in the task and shirts. The experimental results show that the proposed method can monitor processes in domains where state representation is inherently challenging.}},
  author       = {{Verleysen, Andreas and Biondina, Matthijs and wyffels, Francis}},
  issn         = {{0924-669X}},
  journal      = {{APPLIED INTELLIGENCE}},
  keywords     = {{Artificial Intelligence,Process monitoring,Deformable object manipulation,Contrastive learning,Learning from demonstrations,Semantic representation}},
  language     = {{eng}},
  number       = {{2}},
  pages        = {{1725--1743}},
  title        = {{Learning self-supervised task progression metrics : a case of cloth folding}},
  url          = {{http://doi.org/10.1007/s10489-022-03466-8}},
  volume       = {{53}},
  year         = {{2023}},
}
Altmetric
View in Altmetric
Web of Science
Times cited: