Pattern or artifact? Interactively exploring embedding quality with TRACE
- Author
- Edith Heiter (UGent) , Liesbet Martens (UGent) , Ruth Seurinck (UGent) , Martin Guilliams (UGent) , Tijl De Bie (UGent) , Yvan Saeys (UGent) and Jefrey Lijffijt (UGent)
- Organization
- Project
-
- Automating Data Science: the Next Frontiers
- Flanders Artificial Intelligence Research program (FAIR) – second cycle - 2024
- Interactive embedding of high-dimensional data with shape templates
- Exploring Data: Theoretical Foundations and Applications to Web, Multimedia, and Omics Data
- Conditional Knowledge Graph Embedding
- The activation of inactive people: studying thresholds and breakthroughs using AI techniques
- Abstract
- This paper presents TRACE, a tool to analyze the quality of 2D embeddings generated through dimensionality reduction techniques. Dimensionality reduction methods often prioritize preserving either local neighborhoods or global distances, but insights from visual structures can be misleading if the objective has not been achieved uniformly. TRACE addresses this challenge by providing a scalable and extensible pipeline for computing both local and global quality measures. The interactive browser-based interface allows users to explore various embeddings while visually assessing the pointwise embedding quality. The interface also facilitates in-depth analysis by highlighting high-dimensional nearest neighbors for any group of points and displaying high-dimensional distances between points. TRACE enables analysts to make informed decisions regarding the most suitable dimensionality reduction method for their specific use case, by showing the degree and location where structure is preserved in the reduced space.
- Keywords
- Visualisation, Interactive visualisation, Visual analytics, Dimensionality Reduction, Evaluation
Downloads
-
TRACE accepted.pdf
- full text (Accepted manuscript)
- |
- open access
- |
- |
- 749.82 KB
-
(...).pdf
- full text (Published version)
- |
- UGent only
- |
- |
- 1.38 MB
Citation
Please use this url to cite or link to this publication: http://hdl.handle.net/1854/LU-01J6S0DQS4X2SX0BT15PTXHY97
- MLA
- Heiter, Edith, et al. “Pattern or Artifact? Interactively Exploring Embedding Quality with TRACE.” MACHINE LEARNING AND KNOWLEDGE DISCOVERY IN DATABASES-RESEARCH TRACK AND DEMO TRACK, PT VIII, ECML PKDD 2024, vol. 14948, Springer Cham, 2024, pp. 379–82, doi:10.1007/978-3-031-70371-3_24.
- APA
- Heiter, E., Martens, L., Seurinck, R., Guilliams, M., De Bie, T., Saeys, Y., & Lijffijt, J. (2024). Pattern or artifact? Interactively exploring embedding quality with TRACE. MACHINE LEARNING AND KNOWLEDGE DISCOVERY IN DATABASES-RESEARCH TRACK AND DEMO TRACK, PT VIII, ECML PKDD 2024, 14948, 379–382. https://doi.org/10.1007/978-3-031-70371-3_24
- Chicago author-date
- Heiter, Edith, Liesbet Martens, Ruth Seurinck, Martin Guilliams, Tijl De Bie, Yvan Saeys, and Jefrey Lijffijt. 2024. “Pattern or Artifact? Interactively Exploring Embedding Quality with TRACE.” In MACHINE LEARNING AND KNOWLEDGE DISCOVERY IN DATABASES-RESEARCH TRACK AND DEMO TRACK, PT VIII, ECML PKDD 2024, 14948:379–82. Springer Cham. https://doi.org/10.1007/978-3-031-70371-3_24.
- Chicago author-date (all authors)
- Heiter, Edith, Liesbet Martens, Ruth Seurinck, Martin Guilliams, Tijl De Bie, Yvan Saeys, and Jefrey Lijffijt. 2024. “Pattern or Artifact? Interactively Exploring Embedding Quality with TRACE.” In MACHINE LEARNING AND KNOWLEDGE DISCOVERY IN DATABASES-RESEARCH TRACK AND DEMO TRACK, PT VIII, ECML PKDD 2024, 14948:379–382. Springer Cham. doi:10.1007/978-3-031-70371-3_24.
- Vancouver
- 1.Heiter E, Martens L, Seurinck R, Guilliams M, De Bie T, Saeys Y, et al. Pattern or artifact? Interactively exploring embedding quality with TRACE. In: MACHINE LEARNING AND KNOWLEDGE DISCOVERY IN DATABASES-RESEARCH TRACK AND DEMO TRACK, PT VIII, ECML PKDD 2024. Springer Cham; 2024. p. 379–82.
- IEEE
- [1]E. Heiter et al., “Pattern or artifact? Interactively exploring embedding quality with TRACE,” in MACHINE LEARNING AND KNOWLEDGE DISCOVERY IN DATABASES-RESEARCH TRACK AND DEMO TRACK, PT VIII, ECML PKDD 2024, Vilnius, Lithuania, 2024, vol. 14948, pp. 379–382.
@inproceedings{01J6S0DQS4X2SX0BT15PTXHY97,
abstract = {{This paper presents TRACE, a tool to analyze the quality of 2D embeddings generated through dimensionality reduction techniques. Dimensionality reduction methods often prioritize preserving either local neighborhoods or global distances, but insights from visual structures can be misleading if the objective has not been achieved uniformly. TRACE addresses this challenge by providing a scalable and extensible pipeline for computing both local and global quality measures. The interactive browser-based interface allows users to explore various embeddings while visually assessing the pointwise embedding quality. The interface also facilitates in-depth analysis by highlighting high-dimensional nearest neighbors for any group of points and displaying high-dimensional distances between points. TRACE enables analysts to make informed decisions regarding the most suitable dimensionality reduction method for their specific use case, by showing the degree and location where structure is preserved in the reduced space.}},
author = {{Heiter, Edith and Martens, Liesbet and Seurinck, Ruth and Guilliams, Martin and De Bie, Tijl and Saeys, Yvan and Lijffijt, Jefrey}},
booktitle = {{MACHINE LEARNING AND KNOWLEDGE DISCOVERY IN DATABASES-RESEARCH TRACK AND DEMO TRACK, PT VIII, ECML PKDD 2024}},
isbn = {{9783031703706}},
issn = {{0302-9743}},
keywords = {{Visualisation,Interactive visualisation,Visual analytics,Dimensionality Reduction,Evaluation}},
language = {{eng}},
location = {{Vilnius, Lithuania}},
pages = {{379--382}},
publisher = {{Springer Cham}},
title = {{Pattern or artifact? Interactively exploring embedding quality with TRACE}},
url = {{http://doi.org/10.1007/978-3-031-70371-3_24}},
volume = {{14948}},
year = {{2024}},
}
- Altmetric
- View in Altmetric
- Web of Science
- Times cited: