Microsoft Research

Humans can easily see 3D shape from single 2D images, exploiting multiple kinds of information.

StreamSVM is a solver for Linear SVMs that exploits the different speeds of computing on the CPU and accessing data from disk.

Positive definite matrices arise in several contexts: quantum mechanics, statistics, machine learning, image processing, elasticity etc. Various notions of averaging are useful in different contexts.

Many applications of statistical learning involve estimating a sparse covariance matrix from a sample set of random variables.

Microsoft Research – 12 августа 2016, 2:07

For more than 75 years, elections forecasting has been staticΓÇöask a random sample of a representative group of voters who they would vote for if the election were held today, and then report the

Bayesian approaches to clustering permit great flexibility existing models can handle cases when the number of clusters is not known upfront, or when one wants to share clusters across multiple data

Microsoft Research – 12 августа 2016, 2:07

Juan Vargas from Microsoft chairs this session at Faculty Summit 2012. This session continues with the theme ΓÇ£The Changing Landscape of Parallel ComputingΓÇ¥.

I will discuss creating (conjectural) tables of elliptic curves over Q(√5) ordered by conductor up to the first curve of rank 2.

11:00ΓÇô12:00: Tools (Testing and Debugging) (UIUC 30 mins) (UCB 15 mins)(MSFT 15 mins) Darko Marinov (UIUC): Immunity + Pennelope (15 mins) Dany Dig (UIUC): Refactoring (15 mins) Koushik Sen(UCB)

Rane Johnson from Microsoft Research chairs this session at Faculty Summit 2012.

Consider a dataset of vector-valued observations that consists of a modest number of noisy inliers, which are explained well by a low-dimensional subspace, along with a large number of outliers

The k-means algorithm is still the most popular clustering algorithm in practice. Even though it is known to perform very well in practice, it does not give any approximation guarantees.

I will give an introduction to the basic multi-armed bandit problem, mention some of its many variants, and time permitting give an indication of proof techniques.

Microsoft Research – 12 августа 2016, 2:05

Microsoft Research – 12 августа 2016, 2:05

danah boyd from Microsoft Research chairs this session at Faculty Summit 2012.

SPEAKER: Kiran Kedlaya TITLE: The Robbins phenomenon: unexpected numerical stability in p-adic arithmetic ABSTRACT: Since one cannot represent an arbitrary real number on a computer, it is standard

9:00-10:30: Programming Systems David Padua (UIUC): Abstractions for Parallel Programming (15 mins) Armando Fox, Shoaib Kamil (UC Berkeley): SEJITS (30 mins) David Callahan (Microsoft): C++ AMP (30

As corporations, agencies, and individuals continue to invest in national infrastructure trusting it to withstand cyber-attacks, it is important to ensure that the this trust is warranted.

We show how certain random projections and random sampling methods can be used to design efficient dimensionality reduction techniques for two popular machine learning problems: (i) K-means

Principal Component Analysis is one of the most widely used techniques for dimensionality reduction. Nevertheless, it is plagued by sensitivity to outliers; finding robust analogs is critical.

Donald Brinkman from Microsoft Research chairs this session at Faculty Summit 2012.

We will discuss a new approach to proving the Ferrero-Greenberg formula for the derivative of a Kubota-Leoplodt $p$-adic $L$-function at $s=0$.

Juan Vargas from Microsoft chairs this session at Faculty Summit 2012.

3:30-4:30: Architecture (Intel, UIUC, UCB (20 mins each) Joseph Torrellas (UIUC): The Bulk Architecture (20 mins) Krste Asanovic (UCB): ΓÇ£21st Century Computer Architecture ResearchΓÇ¥ Tim Mattson

4:30ΓÇô5:30: Panel: UPCRC; Can Industry & Academia Collaborations be Effective? Dave Patterson (UCB) Burton Smith, Jim Larus (Microsoft) Tim Mattson (Intel) Josep Torrellas (UIUC)

Categories in multi-class data are often part of an underlying semantic taxonomy.

Microsoft Research – 12 августа 2016, 2:01