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

To facilitate a wide-spread acceptance of AI systems guiding decision makingin real-world applications, trustworthiness of deployed models is key. That is,it is crucial for predictive models to be uncertainty-aware and yieldwell-calibrated (and thus trustworthy) predictions for both in-domain samplesas well as under domain shift. Recent efforts to account for predictiveuncertainty include post-processing steps for trained neural networks, Bayesianneural networks as well as alternative non-Bayesian approaches such as ensembleapproaches and evidential deep learning. Here, we propose an efficient yetgeneral modelling approach for obtaining well-calibrated, trustworthyprobabilities for samples obtained after a domain shift. We introduce a newtraining strategy combining an entropy-encouraging loss term with anadversarial calibration loss term and demonstrate that this results inwell-calibrated and technically trustworthy predictions for a wide range ofdomain drifts. We comprehensively evaluate previously proposed approaches ondifferent data modalities, a large range of data sets including sequence data,network architectures and perturbation strategies. We observe that ourmodelling approach substantially outperforms existing state-of-the-artapproaches, yielding well-calibrated predictions under domain drift.

Towards Trustworthy Predictions from Deep Neural Networks with Fast Adversarial Calibration