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

$\textbf{Background:}$ At the onset of a pandemic, such as COVID-19, datawith proper labeling/attributes corresponding to the new disease might beunavailable or sparse. Machine Learning (ML) models trained with the availabledata, which is limited in quantity and poor in diversity, will often be biasedand inaccurate. At the same time, ML algorithms designed to fight pandemicsmust have good performance and be developed in a time-sensitive manner. Totackle the challenges of limited data, and label scarcity in the availabledata, we propose generating conditional synthetic data, to be used alongsidereal data for developing robust ML models. $\textbf{Methods:}$ We present ahybrid model consisting of a conditional generative flow and a classifier forconditional synthetic data generation. The classifier decouples the featurerepresentation for the condition, which is fed to the flow to extract the localnoise. We generate synthetic data by manipulating the local noise with fixedconditional feature representation. We also propose a semi-supervised approachto generate synthetic samples in the absence of labels for a majority of theavailable data. $\textbf{Results:}$ We performed conditional syntheticgeneration for chest computed tomography (CT) scans corresponding to normal,COVID-19, and pneumonia afflicted patients. We show that our methodsignificantly outperforms existing models both on qualitative and quantitativeperformance, and our semi-supervised approach can efficiently synthesizeconditional samples under label scarcity. As an example of downstream use ofsynthetic data, we show improvement in COVID-19 detection from CT scans withconditional synthetic data augmentation.

Conditional Synthetic Data Generation for Robust Machine Learning Applications with Limited Pandemic Data