Microsoft Research

Characterizing the computational complexity of statistical inference problems is an outstanding open problem.

The objective in extreme multi-label classification is to learn a classifier that can automatically tag a data point with the most relevant subset of labels from a large label set.

In a classical problem in scheduling, one has n unit size jobs with a precedence order and the goal is to find a schedule of those jobs on m identical machines as to minimize the makespan.

We present data-dependent learning bounds for the general scenario of non-stationary non-mixing stochastic processes.

In the past several years, there has been a lot of progress on combinatorial optimization.

Type a summary that describes the content of your video.

Consider a branching tree with one root at the origin. Assuming that at the n -th level there are at least two vertices. Pick two of them by simple random sampling without replacement.

Since being analyzed by Rokhlin, Szlam, and Tygert and popularized by Halko, Martinsson, and Tropp, randomized Simultaneous Power Iteration has become the method of choice for approximate singular

O(T^{-3/4}) O(T^{-1}) O(T^{-1/2}) \tilde{O}(T^{-1/2}) We show that natural classes of regularized learning algorithms with a form of recency bias achieve faster convergence rates to approximate

We will see the basics of Ito calculus and Girsanov's change of measure formula.

In recent years, Sparse Principal Component Analysis has emerged as an extremely popular dimension reduction technique for high-dimensional data.

Writing papers and giving talks are key skills for any researcher, but they aren’t easy.

We consider a setting for nonstochastic multiarmed bandits in which actions are vertices of a graph G, the edges of G denote similarities between actions, and the payoffs observed in each play are

Monte Carlo methods use random sampling to understand a system, estimate averages, or compute integrals.

Deep Learning algorithms attempt to discover good representations, at multiple levels of abstraction.

Welcome, Introductions, & Short Talk Ideas

No Abstract Available

We introduce a neural network with a recurrent attention model over a possibly large external memory.

A distributed algorithm for computing the majority between two populations

A short introduction to the Infer.NET Bayesian inference framework.

Why I work at Microsoft Research - Jonathan Taylor, Researcher, Machine Intelligence and Perception

The Great Divide: Unequal Societies and What We Can Do About Them In The Great Divide, Joseph E.

Perhaps the most important reason eye tracking will play an ever more crucial role in human computer interactions is that our eye activity is driven by subconscious processes in our brain.

Today’s Science, Tomorrow’s Engineering The mission of CBMM is to make progress on the greatest problem in science — human intelligence.

Decades ago, when the audio industry was in its infancy, there was no way to get formal education in Audio Engineering.

Video of animated graphics projected onto the trunk of a fiber optic tree.

In the field of computer networks, and undoubtedly in other engineering fields, a lot of what "everybody knows" is ... not true.

The frog model is a branching random walk that splits into two particles upon the first visit to each site. We prove a phase transition from recurrence to transience on the rooted d-ary tree.

Social Computing Symposium 2015: Welcome and Introductions

Panel Discussion: Interdisciplinarity: The Future of Computer Science?