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

Deep neural networks have been successful in diverse discriminativeclassification tasks, although, they are poorly calibrated often assigning highprobability to misclassified predictions. Potential consequences could lead totrustworthiness and accountability of the models when deployed in realapplications, where predictions are evaluated based on their confidence scores.Existing solutions suggest the benefits attained by combining deep neuralnetworks and Bayesian inference to quantify uncertainty over the models'predictions for ambiguous datapoints. In this work we propose to validate andtest the efficacy of likelihood based models in the task of out of distributiondetection (OoD). Across different datasets and metrics we show that Bayesiandeep learning models on certain occasions marginally outperform conventionalneural networks and in the event of minimal overlap between in/out distributionclasses, even the best models exhibit a reduction in AUC scores in detectingOoD data. Preliminary investigations indicate the potential inherent role ofbias due to choices of initialisation, architecture or activation functions. Wehypothesise that the sensitivity of neural networks to unseen inputs could be amulti-factor phenomenon arising from the different architectural design choicesoften amplified by the curse of dimensionality. Furthermore, we perform a studyto find the effect of the adversarial noise resistance methods on in andout-of-distribution performance, as well as, also investigate adversarial noiserobustness of Bayesian deep learners.

Ramifications of Approximate Posterior Inference for Bayesian Deep Learning in Adversarial and Out-of-Distribution Settings