Michael I. Jordan

TL;DR: A family of efficient algorithms that use EM and the Minimum Spanning Tree algorithm to find the ML and MAP mixture of trees for a variety of priors, including the Dirichlet and the MDL priors are presented.

TL;DR: In this article, a fully Bayesian framework for integrating discrete mixed membership and continuous latent factor models into unified Mixed Membership Matrix Factorization (M3F) models is developed, and two M3F models, derived Gibbs sampling inference procedures, are introduced and validated on the EachMovie, MovieLens, and Netflix Prize collaborative filtering datasets.

TL;DR: In this article, a generalization of the clustering problem, called feature allocation, is proposed, where each data point can belong to an arbitrary, non-negative integer number of groups, now called features or topics.

TL;DR: In this article, the authors provide a detailed study of the estimation of probability distributions in a stringent setting in which data is kept private even from the statistician, and give sharp minimax rates of convergence for estimation in these locally private settings, exhibiting fundamental trade-offs between privacy and convergence rate.

TL;DR: This paper defines such priors as a mixture of exponential power distributions with a generalized inverse Gaussian density (EP-GIG), a variant of generalized hyperbolic distributions, and shows that these algorithms bear an interesting resemblance to iteratively reweighted l2 or l1 methods.