It is now well established that sparse signal models are well suited for restoration tasks and can be effectively learned from audio, image, and video data. Recent research has be...
Julien Mairal, Francis Bach, Jean Ponce, Guillermo...
Graph-based semi-supervised learning has recently emerged as a promising approach to data-sparse learning problems in natural language processing. All graph-based algorithms rely ...
Learning tasks from a single demonstration presents a significant challenge because the observed sequence is inherently an incomplete representation of the procedure that is speci...
Hyuckchul Jung, James F. Allen, Nathanael Chambers...
Recently, relevance vector machines (RVM) have been fashioned from a sparse Bayesian learning (SBL) framework to perform supervised learning using a weight prior that encourages s...
Manifold learning can discover the structure of high dimensional data and provides understanding of multidimensional patterns by preserving the local geometric characteristics. Ho...