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

Classical problems in computational physics such as data-driven forecastingand signal reconstruction from sparse sensors have recently seen an explosionin deep neural network (DNN) based algorithmic approaches. However, most DNNmodels do not provide uncertainty estimates, which are crucial for establishingthe trustworthiness of these techniques in downstream decision making tasks andscenarios. In recent years, ensemble-based methods have achieved significantsuccess for the uncertainty quantification in DNNs on a number of benchmarkproblems. However, their performance on real-world applications remainsunder-explored. In this work, we present an automated approach to DNN discoveryand demonstrate how this may also be utilized for ensemble-based uncertaintyquantification. Specifically, we propose the use of a scalable neural andhyperparameter architecture search for discovering an ensemble of DNN modelsfor complex dynamical systems. We highlight how the proposed method not onlydiscovers high-performing neural network ensembles for our tasks, but alsoquantifies uncertainty seamlessly. This is achieved by using genetic algorithmsand Bayesian optimization for sampling the search space of neural networkarchitectures and hyperparameters. Subsequently, a model selection approach isused to identify candidate models for an ensemble set construction. Afterwards,a variance decomposition approach is used to estimate the uncertainty of thepredictions from the ensemble. We demonstrate the feasibility of this frameworkfor two tasks - forecasting from historical data and flow reconstruction fromsparse sensors for the sea-surface temperature. We demonstrate superiorperformance from the ensemble in contrast with individual high-performingmodels and other benchmarks.

Quantifying uncertainty for deep learning based forecasting and flow-reconstruction using neural architecture search ensembles