English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Concept discovery is one of the open problems in the interpretabilityliterature that is important for bridging the gap between non-deep learningexperts and model end-users. Among current formulations, concepts defines themby as a direction in a learned representation space. This definition makes itpossible to evaluate whether a particular concept significantly influencesclassification decisions for classes of interest. However, finding relevantconcepts is tedious, as representation spaces are high-dimensional and hard tonavigate. Current approaches include hand-crafting concept datasets and thenconverting them to latent space directions; alternatively, the process can beautomated by clustering the latent space. In this study, we offer another twoapproaches to guide user discovery of meaningful concepts, one based onmultiple hypothesis testing, and another on interactive visualization. Weexplore the potential value and limitations of these approaches throughsimulation experiments and an demo visual interface to real data. Overall, wefind that these techniques offer a promising strategy for discovering relevantconcepts in settings where users do not have predefined descriptions of them,but without completely automating the process.

Discovering Concepts in Learned Representations using Statistical Inference and Interactive Visualization