Image 1_A multimodal AI-driven framework for cardiovascular screening and risk assessment in diverse athletic populations: innovations in sports cardiology.png by Minjin Guo (22751300)

“…</p>Methods<p>To address these challenges, we propose a novel AI-driven framework that incorporates two key methodological innovations: CardioSpectra, a structured sparse inference model, and Risk-Stratified Exertional Embedding (RSEE), a domain-specific representation learning strategy. CardioSpectra formulates athlete profiles as multivariate probabilistic entities across latent diagnostic states, using sparsity-aware inference to generate interpretable risk predictions while optimizing a sensitivity-specificity trade-off tailored to clinical priorities. …”