research.draconitic.com

Monday 16 November

Domain-decoupled physics-informed neural networks for inverse material property estimation from acoustic… — Draconitic Research

AI-generated summary Draconitic Research describes a method using sequential physics-informed neural networks with Fourier features to estimate material properties for a vibrating plate from radiated acoustic fields. The approach decouples fluid and structural domains to filter boundary differentiation and data noise while introducing multi-frequency statistical ensembles to quantify uncertainty.

Sunday 15 November

Carotid CTA with automated plaque quantification for risk stratification of coexisting obstructive coronary… — Draconitic Research

AI-generated summary Draconitic Research evaluated whether automated carotid plaque quantification on carotid computed tomography angiography can serve as an adjunctive risk stratification tool for identifying patients with coexisting obstructive coronary artery disease. In a retrospective study including 2533 adults, calcified plaque volume and fibrotic plaque volume emerged as independent predictors associated with obstructive coronary artery disease.