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

The rise of deep learning has ushered in significant progress in computervision (CV) tasks, yet the "black box" nature of these models often precludesinterpretability. This challenge has spurred the development of ExplainableArtificial Intelligence (XAI) by generating explanations to AI'sdecision-making process. An explanation is aimed to not only faithfully reflectthe true reasoning process (i.e., faithfulness) but also align with humans'reasoning (i.e., alignment). Within XAI, visual explanations employ visual cuesto elucidate the reasoning behind machine learning models, particularly inimage processing, by highlighting images' critical areas important topredictions. Despite the considerable body of research in visual explanations,standardized benchmarks for evaluating them are seriously underdeveloped. Inparticular, to evaluate alignment, existing works usually merely illustrate afew images' visual explanations, or hire some referees to report theexplanation quality under ad-hoc questionnaires. However, this cannot achieve astandardized, quantitative, and comprehensive evaluation. To address thisissue, we develop a benchmark for visual explanation, consisting of eightdatasets with human explanation annotations from various domains, accommodatingboth post-hoc and intrinsic visual explanation methods. Additionally, we devisea visual explanation pipeline that includes data loading, explanationgeneration, and method evaluation. Our proposed benchmarks facilitate a fairevaluation and comparison of visual explanation methods. Building on ourcurated collection of datasets, we benchmarked eight existing visualexplanation methods and conducted a thorough comparison across four selecteddatasets using six alignment-based and causality-based metrics. Our benchmarkwill be accessible through our website https://xaidataset.github.io.

XAI Benchmark for Visual Explanation