On connecting sparsity and deep learning
Sparsity plays a pivotal role in statistics and machine learning. In this talk we discuss how a generalized and hierarchical view of sparsity may be used to explain the underpinnings of deep learning. This provides a rigorous statistical foundation for deep learning, which has demonstrated significant success in practice. We demonstrate applications of these ideas at the intersection of machine learning and biological sciences.
Comment
Your name
Ted Dick, Case Western Reserve University
Table sorting checkbox
Off
Is there an actionable arm to your model to identify effectve interventions or is that determined by the physician? Can you use this approach to assess the effect of interventions? Is there a temporal compenent in which the clustering of words deends on date of entry?