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

Deep Learning of neural networks has gained prominence in multiplelife-critical applications like medical diagnoses and autonomous vehicleaccident investigations. However, concerns about model transparency and biasespersist. Explainable methods are viewed as the solution to address thesechallenges. In this study, we introduce the Occlusion Sensitivity Analysis withDeep Feature Augmentation Subspace (OSA-DAS), a novel perturbation-basedinterpretability approach for computer vision. While traditional perturbationmethods make only use of occlusions to explain the model predictions, OSA-DASextends standard occlusion sensitivity analysis by enabling the integrationwith diverse image augmentations. Distinctly, our method utilizes the outputvector of a DNN to build low-dimensional subspaces within the deep featurevector space, offering a more precise explanation of the model prediction. Thestructural similarity between these subspaces encompasses the influence ofdiverse augmentations and occlusions. We test extensively on the ImageNet-1k,and our class- and model-agnostic approach outperforms commonly usedinterpreters, setting it apart in the realm of explainable AI.

Occlusion Sensitivity Analysis with Augmentation Subspace Perturbation in Deep Feature Space