Microsoft Research

A simple random walk (SRW) on a graph is a Markov chain whose state space is the vertex set and the next state distribution is uniform among the neighbors of the current state.

Data sets usually have some structural property that make them useful. An array of points may be sorted, a set of points may be in convex position, a graph may be a tree, etc.

A common approach in designing parallel languages is to provide some high level handles to manipulate the use of the parallel platform.

There are a range of human reactions to music: from indifference to rapture, from ΓÇ£amusiaΓÇ¥ (the inability to recognize certain aspects of music) to a synesthetic response that imbues every

Optimal control in biology and engineering

Microsoft Research – 6 сентября 2016, 16:52

Microsoft Research – 6 сентября 2016, 16:52

One of the great landmarks in the history of science occurred when J.

Microsoft Research – 6 сентября 2016, 16:51

We study particle systems corresponding to highly connected queuing networks.

Current climate models work by attempting to simulate the weather from first principles.

Small companies form a large part of the software industry, but have mostly been overlooked by the requirements engineering research community.

The tremendous increase in available biological information creates opportunities to decipher the structure, function and evolution of cellular components while presenting new computational

In recent years there has been considerable interest in analyzing random graph models for the Web. We consider two such models - the Random Surfer model introduced by Blum et al.

Microsoft Research – 6 сентября 2016, 16:49

Resource-Based NLP (by Hitoshi Isahara) Abstract: At NICT, we develop widely-applicable and high-performance NLP technologies and linguistic resources, which we make available to public to

In this study, a new discriminative learning framework, called soft margin estimation (SME), is proposed for estimating the parameters of continuous density hidden Markov models (HMMs).

Microsoft Research – 6 сентября 2016, 16:49

Microsoft Research – 6 сентября 2016, 16:49

Microsoft Research – 6 сентября 2016, 16:49

Microsoft Research – 6 сентября 2016, 16:48

Microsoft Research – 6 сентября 2016, 16:48

Microsoft Research – 6 сентября 2016, 16:48

Microsoft Research – 6 сентября 2016, 16:48

Systems code is prone to termination errors. Inferring preconditions excluding such errors is both tedious and hard. This is true in particular for heap manipulating functions.

In the talk, I will introduce an efficient statistical machine learning approach for classifying networked data, stacked graphical learning.

Microsoft Research – 6 сентября 2016, 16:48

In many algorithmic problems like nearest neighbor search, clustering, near-duplicate detection or decentralized navigation input dataset is described either by some explicit representation (e.g.

The statistical learning theory suggests to choose large capacity models that barely avoid over-fitting the training data. In that perspective, all datasets are small.

In this talk I'll draw on work from my visit to Microsoft Cambridge earlier this year, in collaboration with the SDS group.