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

This paper presents a novel evaluation framework for Out-of-Distribution(OOD) detection that aims to assess the performance of machine learning modelsin more realistic settings. We observed that the real-world requirements fortesting OOD detection methods are not satisfied by the current testingprotocols. They usually encourage methods to have a strong bias towards a lowlevel of diversity in normal data. To address this limitation, we propose newOOD test datasets (CIFAR-10-R, CIFAR-100-R, and ImageNet-30-R) that can allowresearchers to benchmark OOD detection performance under realistic distributionshifts. Additionally, we introduce a Generalizability Score (GS) to measure thegeneralization ability of a model during OOD detection. Our experimentsdemonstrate that improving the performance on existing benchmark datasets doesnot necessarily improve the usability of OOD detection models in real-worldscenarios. While leveraging deep pre-trained features has been identified as apromising avenue for OOD detection research, our experiments show thatstate-of-the-art pre-trained models tested on our proposed datasets suffer asignificant drop in performance. To address this issue, we propose apost-processing stage for adapting pre-trained features under thesedistribution shifts before calculating the OOD scores, which significantlyenhances the performance of state-of-the-art pre-trained models on ourbenchmarks.

Towards Realistic Out-of-Distribution Detection: A Novel Evaluation Framework for Improving Generalization in OOD Detection