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

Calibration of neural networks is a topical problem that is becoming more andmore important as neural networks increasingly underpin real-worldapplications. The problem is especially noticeable when using modern neuralnetworks, for which there is a significant difference between the confidence ofthe model and the probability of correct prediction. Various strategies havebeen proposed to improve calibration, yet accurate calibration remainschallenging. We propose a novel framework with two contributions: introducing anew differentiable surrogate for expected calibration error (DECE) that allowscalibration quality to be directly optimised, and a meta-learning frameworkthat uses DECE to optimise for validation set calibration with respect to modelhyper-parameters. The results show that we achieve competitive performance withexisting calibration approaches. Our framework opens up a new avenue andtoolset for tackling calibration, which we believe will inspire further work onthis important challenge.

Meta-Calibration: Learning of Model Calibration Using Differentiable Expected Calibration Error