The Semantic Librarian project creates search engines using vector-space models of semantics. It converts words, sentences, and documents to points in a high-dimensional meaning space, where texts that are closer together are more similar in their semantic meaning.

Example

A major aim of the project is to create tools for semantic space visualization. The example shows a recent deployment of the Semantic Librarian for conference program abstracts. Users use semantic similarities between search terms, documents, and document authors to find conference abstracts that may be of interest to them. The Shiny app UI presents a multi-dimensional scaling solution showing the similarity space of returned items, as well as a table of documents ranked by similarity.

Browse

We have developed a few different Semantic Librarians, which are currently hosted as Shiny apps. You can explore them here.

APA

American Psychological Association journal abstracts from 1890s-2016.

CSBBCS

Conference abstracts for Canadian Society for Brain and Behavioural Sciences 2019 meeting.

SCiP

Conference abstracts for Society for Computers in Psychology (last 10 years)

Psychonomics 2019

Conference abstracts for Psychonomics 2019.