By Username
Enter a URL of an MRSS feed

Techniques for Massive-Data Machine Learning, with Application to Astronomy We'll begin by looking at seven fundamental problems in astronomy, and the state-of-the-art methods methods that can be used to solve them, which include n-point correlation functions, kernel density estimation, minimum spanning trees, bipartite matching, nonparametric Bayes classifiers, support vector machines, Nadaraya-Watson regression, kernel conditional density estimation, Gaussian process regression, nearest-neighbors, principal component analysis, hierarchical clustering, and manifold learning. I will then describe how we can scale each of these methods to work on massive survey datasets, despite their often quadratic or cubic scaling with the number of data, via seven different types of computational techniques: indexing, functional transforms, sampling, problem reductions, locality, parallelism, and active learning. See other lectures at Purdue MLSS Playlist: