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

How do we know when the predictions made by a classifier can be trusted? Thisis a fundamental problem that also has immense practical applicability,especially in safety-critical areas such as medicine and autonomous driving.The de facto approach of using the classifier's softmax outputs as a proxy fortrustworthiness suffers from the over-confidence issue; while the most recentworks incur problems such as additional retraining cost and accuracy versustrustworthiness trade-off. In this work, we argue that the trustworthiness of aclassifier's prediction for a sample is highly associated with two factors: thesample's neighborhood information and the classifier's output. To combine thebest of both worlds, we design a model-agnostic post-hoc approach NeighborAggto leverage the two essential information via an adaptive neighborhoodaggregation. Theoretically, we show that NeighborAgg is a generalized versionof a one-hop graph convolutional network, inheriting the powerful modelingability to capture the varying similarity between samples within each class. Wealso extend our approach to the closely related task of mislabel detection andprovide a theoretical coverage guarantee to bound the false negative.Empirically, extensive experiments on image and tabular benchmarks verify ourtheory and suggest that NeighborAgg outperforms other methods, achievingstate-of-the-art trustworthiness performance.

Birds of a Feather Trust Together: Knowing When to Trust a Classifier via Adaptive Neighborhood Aggregation