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

Using deep learning (DL) to accelerate and/or improve scientific workflowscan yield discoveries that are otherwise impossible. Unfortunately, DL modelshave yielded limited success in complex scientific domains due to large datarequirements. In this work, we propose to overcome this issue by integratingthe abundance of scientific knowledge sources (SKS) with the DL trainingprocess. Existing knowledge integration approaches are limited to usingdifferentiable knowledge source to be compatible with first-order DL trainingparadigm. In contrast, our proposed approach treats knowledge source as ablack-box in turn allowing to integrate virtually any knowledge source. Toenable an end-to-end training of SKS-coupled-DL, we propose to use zeroth-orderoptimization (ZOO) based gradient-free training schemes, which isnon-intrusive, i.e., does not require making any changes to the SKS. Weevaluate the performance of our ZOO training scheme on two real-world materialscience applications. We show that proposed scheme is able to effectivelyintegrate scientific knowledge with DL training and is able to outperformpurely data-driven model for data-limited scientific applications. We alsodiscuss some limitations of the proposed method and mention potentiallyworthwhile future directions.

Zeroth-Order SciML: Non-intrusive Integration of Scientific Software with Deep Learning