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

Prototypical methods have recently gained a lot of attention due to theirintrinsic interpretable nature, which is obtained through the prototypes. Withgrowing use cases of model reuse and distillation, there is a need to alsostudy transfer of interpretability from one model to another. We presentProto2Proto, a novel method to transfer interpretability of one prototypicalpart network to another via knowledge distillation. Our approach aims to addinterpretability to the "dark" knowledge transferred from the teacher to theshallower student model. We propose two novel losses: "Global Explanation" lossand "Patch-Prototype Correspondence" loss to facilitate such a transfer. GlobalExplanation loss forces the student prototypes to be close to teacherprototypes, and Patch-Prototype Correspondence loss enforces the localrepresentations of the student to be similar to that of the teacher. Further,we propose three novel metrics to evaluate the student's proximity to theteacher as measures of interpretability transfer in our settings. Wequalitatively and quantitatively demonstrate the effectiveness of our method onCUB-200-2011 and Stanford Cars datasets. Our experiments show that the proposedmethod indeed achieves interpretability transfer from teacher to student whilesimultaneously exhibiting competitive performance.

Proto2Proto: Can you recognize the car, the way I do?