Browsing by Subject "Unsupervised machine learning"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    ML.* MACHINE LEARNING LIBRARY AS A MUSICAL PARTNER IN THE COMPUTER-ACOUSTIC COMPOSITION FLIGHT
    (Michigan Publishing, 2014-09) Smith, Benjamin D.; Deal, W. Scott
    This paper presents an application and extension of the ml.* library, implementing machine learning (ML) models to facilitate “creative” interactions between musician and machine. The objective behind the work is to effectuate a musical “virtual partner” capable of creation in a range of musical scenarios that encompass composition, improvisation, studio, and live concert performance. An overview of the piece, Flights, used to test the musical range of the application is given, followed by a description of the development rationale for the project. Its contribution to the aesthetic quality of the human musical process is discussed.
  • Thumbnail Image
    Item
    Unsupervised representation learning on high-dimensional clinical data improves genomic discovery and prediction
    (Springer Nature, 2024) Yun, Taedong; Cosentino, Justin; Behsaz, Babak; McCaw, Zachary R.; Hill, Davin; Luben, Robert; Lai, Dongbing; Bates, John; Yang, Howard; Schwantes-An, Tae-Hwi; Zhou, Yuchen; Khawaja, Anthony P.; Carroll, Andrew; Hobbs, Brian D.; Cho, Michael H.; McLean, Cory Y.; Hormozdiari, Farhad; Medical and Molecular Genetics, School of Medicine
    Although high-dimensional clinical data (HDCD) are increasingly available in biobank-scale datasets, their use for genetic discovery remains challenging. Here we introduce an unsupervised deep learning model, Representation Learning for Genetic Discovery on Low-Dimensional Embeddings (REGLE), for discovering associations between genetic variants and HDCD. REGLE leverages variational autoencoders to compute nonlinear disentangled embeddings of HDCD, which become the inputs to genome-wide association studies (GWAS). REGLE can uncover features not captured by existing expert-defined features and enables the creation of accurate disease-specific polygenic risk scores (PRSs) in datasets with very few labeled data. We apply REGLE to perform GWAS on respiratory and circulatory HDCD-spirograms measuring lung function and photoplethysmograms measuring blood volume changes. REGLE replicates known loci while identifying others not previously detected. REGLE are predictive of overall survival, and PRSs constructed from REGLE loci improve disease prediction across multiple biobanks. Overall, REGLE contain clinically relevant information beyond that captured by existing expert-defined features, leading to improved genetic discovery and disease prediction.