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

There exists an apparent negative correlation between performance andinterpretability of deep learning models. In an effort to reduce this negativecorrelation, we propose a Born Identity Network (BIN), which is a post-hocapproach for producing multi-way counterfactual maps. A counterfactual maptransforms an input sample to be conditioned and classified as a target label,which is similar to how humans process knowledge through counterfactualthinking. For example, a counterfactual map can localize hypotheticalabnormalities from a normal brain image that may cause it to be diagnosed witha disease. Specifically, our proposed BIN consists of two core components:Counterfactual Map Generator and Target Attribution Network. The CounterfactualMap Generator is a variation of conditional GAN which can synthesize acounterfactual map conditioned on an arbitrary target label. The TargetAttribution Network provides adequate assistance for generating synthesizedmaps by conditioning a target label into the Counterfactual Map Generator. Wehave validated our proposed BIN in qualitative and quantitative analysis onMNIST, 3D Shapes, and ADNI datasets, and showed the comprehensibility andfidelity of our method from various ablation studies.

Born Identity Network: Multi-way Counterfactual Map Generation to Explain a Classifier's Decision