We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.HC

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Human-Computer Interaction

Title: Embedding Comparator: Visualizing Differences in Global Structure and Local Neighborhoods via Small Multiples

Abstract: Embeddings mapping high-dimensional discrete input to lower-dimensional continuous vector spaces have been widely adopted in machine learning applications as a way to capture domain semantics. Interviewing 13 embedding users across disciplines, we find comparing embeddings is a key task for deployment or downstream analysis but unfolds in a tedious fashion that poorly supports systematic exploration. In response, we present the Embedding Comparator, an interactive system that presents a global comparison of embedding spaces alongside fine-grained inspection of local neighborhoods. It systematically surfaces points of comparison by computing the similarity of the $k$-nearest neighbors of every embedded object between a pair of spaces. Through case studies across multiple modalities, we demonstrate our system rapidly reveals insights, such as semantic changes following fine-tuning, language changes over time, and differences between seemingly similar models. In evaluations with 15 participants, we find our system accelerates comparisons by shifting from laborious manual specification to browsing and manipulating visualizations.
Comments: Published in IUI 2022; Equal contribution by first two authors
Subjects: Human-Computer Interaction (cs.HC); Computation and Language (cs.CL); Machine Learning (cs.LG)
Cite as: arXiv:1912.04853 [cs.HC]
  (or arXiv:1912.04853v3 [cs.HC] for this version)

Submission history

From: Brandon Carter [view email]
[v1] Tue, 10 Dec 2019 17:46:43 GMT (2775kb,D)
[v2] Sat, 6 Mar 2021 21:28:57 GMT (4133kb,D)
[v3] Fri, 4 Mar 2022 14:51:40 GMT (9630kb,D)

Link back to: arXiv, form interface, contact.