Geometry and Analysis on Digital Data, Emergent Structures and Knowledge Building
Inferential/diffusion geometries on digital data graphs, enable the organization and analysis of empirical data as well as "signal processing" of functions on data.
In particular we will describe various natural multiscale structures on data which enable automatic ontology and language building. Applications to music organization and analysis (based on audio alone ) and medical data search and diagnosis will be described . These developments extend geometries of spectral graph theory,kernel machines and other machine learning tools.
Ronald Coifman, Yale University