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

Explainable machine learning significantly improves the transparency of deepneural networks~(DNN). However, existing work is constrained to explaining thebehavior of individual model predictions, and lacks the ability to transfer theexplanation across various models and tasks. This limitation results inexplaining various tasks being time- and resource-consuming. To address thisproblem, we develop a pre-trained, DNN-based, generic explainer on large-scaleimage datasets, and leverage its transferability to explain various visionmodels for downstream tasks. In particular, the pre-training of genericexplainer focuses on LEarning Transferable Attribution (LETA). The transferableattribution takes advantage of the versatile output of the target backboneencoders to comprehensively encode the essential attribution for explainingvarious downstream tasks. LETA guides the pre-training of the generic explainertowards the transferable attribution, and introduces a rule-based adaptation ofthe transferable attribution for explaining downstream tasks, without the needfor additional training on downstream data. Theoretical analysis demonstratesthat the pre-training of LETA enables minimizing the explanation error boundaligned with the conditional $\mathcal{V}$-information on downstream tasks.Empirical studies involve explaining three different architectures of visionmodels across three diverse downstream datasets. The experiment resultsindicate LETA is effective in explaining these tasks without the need foradditional training on the data of downstream tasks.

LETA: Learning Transferable Attribution for Generic Vision Explainer