Microsoft Research

The composition of a group determines much of its behavior (are people old or young, PhDs or illiterate, artists or scientists?).

One of the major visual complications confronting face recognition is pose variation. It is generally perceived that a part based representation for faces would be more robust to such pose variations.

How do we know a program does what it claims to do? After clustering mined Android apps by their description topics, we identify outliers in each cluster with respect to their API usage.

I present our work on highly-scalable out-of-core techniques for learning well-calibrated Bayesian network classifiers.

Billions of users browse the web on a daily basis, and there are single websites that have reached over one billion user accounts.

Reusable hardware design IPs have become increasingly parameterizable and even taken on the form of fully on-demand generators.

Megaprojects continue to change the face of the world as we know it. Prof.

Printing-press era documents are difficult for OCR systems to transcribe because these documents are extremely noisy. However, the noise originates from processes that are causally understood.

Adaptive networks consist of a collection of agents with local adaptation and learning abilities.

We present recent work on validating satisfiability refutations. Satis?ability (SAT) is considered as one of the most important core technologies in formal verification and related areas.

We all want computers to understand natural language.

Researchers over the past 40 years have developed significant technologies to assist individuals whose bodies are challenged.

In this talk we overview methodology for predicting and analyzing clinical outcomes, especially focusing on neurological disorders, using functional and structural brain imaging data.

NumPy and recently Pandas have made Python ubiquitous for scientific computing and data analytics.

Gap Probabilities for Zeroes of Stationary Gaussian Functions & Rigidity of 3-Colorings of the d-Dimensional Discrete Torus TALK 1: SPEAKER: Naomi Feldheim TITLE: Gap probabilities for zeroes of

Let G be an infinite group with a finite symmetric generating set S. The corresponding Cayley graph on G has an edge between x,y in G if their ratio xy^{-1} is in S.

This talk is about tapping the enormous amount of sensor data that we generate about ourselves every day.

In this talk, we address the problem of falsifying properties of complex cyber-physical sytems that model the interactions between a discrete controller and a continuous physical environment.

Performing gesture recognition without infrastructure has been a long desired feature for seamless human computer interaction.

Much of our routine online activity leaves behind so-called "traces," or electronic records of those actions.

Researchers and software developers require effective performance evaluation. Researchers must evaluate optimizations or measure overhead.

Consider the problem faced by a seller desiring to maximize revenue with m items, in a market with n unit-demand buyers (interested in buying at most one item).

I will talk about some recent results on upper bounding the number of matroids on a ground set of size n. The talk will consist of two parts.

What would you do to help more than 40 scientists, coming together from all over the US for 2 days, extract high-level insights from a pile of loosely-linked spreadsheets?

Most computer architectures, for example, x86, AMD64 and ARMv7 support efficient operations on vectors of data.

Humans can perform many tasks with ease that remain difficult or impossible for computers.

ISHI is a two-week experiment run in the Skype buildings on the Redmond campus. It is designed to foster health & well-being, collaboration, and communication through gameplay.

The objective is to estimate seasonal snow volumes, relative to historical trends and extremes, in snow-dominated mountains that have austere infrastructure, sparse gauging, challenges of

Mathematical logic was developed in an effort to provide formal foundations for mathematics.

Consider a graph G=(V,E) with a weight value w(v) associated with each vertex v. A demand is a pair of vertices (s,t). A subgraph H satisfies the demand if s and t are connected in H.