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 Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Whilst the primary bottleneck to a number of computational workflows was notso long ago limited by processing power, the rise of machine learningtechnologies has resulted in a paradigm shift which places increasing value onissues related to data curation - i.e., data size, quality, bias, format, andcoverage. Increasingly, data-related issues are equally as important as thealgorithmic methods used to process and learn from the data. Here we introducean open source GPU-accelerated neural network (NN) framework for learningreactive potential energy surfaces (PESs), and investigate the use of real-timeinteractive ab initio molecular dynamics in virtual reality (iMD-VR) as a newstrategy for rapidly sampling geometries along reaction pathways which can beused to train NNs to learn reactive PESs. Focussing on hydrogen abstractionreactions of CN radical with isopentane, we compare the performance of NNstrained using iMD-VR data versus NNs trained using a more traditional method,namely molecular dynamics (MD) constrained to sample a predefined grid ofpoints along hydrogen abstraction reaction coordinates. Both the NN trainedusing iMD-VR data and the NN trained using the constrained MD data reproduceimportant qualitative features of the reactive PESs, such as a low and earlybarrier to abstraction. Quantitatively, learning is sensitive to the trainingdataset. Our results show that user-sampled structures obtained with thequantum chemical iMD-VR machinery enable better sampling in the vicinity of theminimum energy path (MEP). As a result, the NN trained on the iMD-VR data doesvery well predicting energies in the vicinity of the MEP, but less wellpredicting energies for 'off-path' structures. The NN trained on theconstrained MD data does better in predicting energies for 'off-path'structures, given that it included a number of such structures in its trainingset.

Training neural nets to learn reactive potential energy surfaces using interactive quantum chemistry in virtual reality