embComp: Visual Interactive Comparison of Vector Embeddings

Florian Heimerl, Christoph Kralj, Torsten Möller, Michael Gleicher

Publications: Contribution to journalArticlePeer Reviewed

Abstract

This paper introduces embComp, a novel approach for comparing two embeddings that capture the similarity between objects, such as word and document embeddings. We survey scenarios where comparing these embedding spaces is useful. From those scenarios, we derive common tasks, introduce visual analysis methods that support these tasks, and combine them into a comprehensive system. One of embComp's central features are overview visualizations that are based on metrics for measuring differences in the local structure around objects. Summarizing these local metrics over the embeddings provides global overviews of similarities and differences. Detail views allow comparison of the local structure around selected objects and relating this local information to the global views. Integrating and connecting all of these components, embComp supports a range of analysis workflows that help understand similarities and differences between embedding spaces. We assess our approach by applying it in several use cases, including understanding corpora differences via word vector embeddings, and understanding algorithmic differences in generating embeddings.
Original languageEnglish
Pages (from-to)2953-2969
Number of pages17
JournalIEEE Transactions on Visualization and Computer Graphics
Volume28
Issue number8
Early online date21 Dec 2020
DOIs
Publication statusPublished - 1 Aug 2022

Austrian Fields of Science 2012

  • 102013 Human-computer interaction

Keywords

  • DIMENSIONALITY REDUCTION
  • Dimensionality reduction
  • EXPLORATION
  • Measurement
  • Object recognition
  • Stress
  • Task analysis
  • Two dimensional displays
  • VISUALIZATION
  • Visual analytics
  • Visualization
  • machine learning
  • vector embeddings
  • visual comparison

Fingerprint

Dive into the research topics of 'embComp: Visual Interactive Comparison of Vector Embeddings'. Together they form a unique fingerprint.

Cite this