Services

Data is a driving force underlying business decisions and processes. Applying leading technologies and analytics to your existing and new data will help you identify new and better ways to serve, produce, optimize and compete. Information overload hasn’t overwhelmed us, it’s creating a new generation of opportunities and advances.

Services

Data Driven Analysis

Leveraging hidden data patterns to inform decisions

Signal Processing

With decades of collective experience working with spatial and temporal data, our expertise in collecting and reconstructing data, signal compression, quality improvement via noise reduction, and extraction of important features will empower you to interpret and make predictions from your data.

Image Analysis

Our extensive experience handling images at all scales from 2D to 4D or higher-dimensional data includes object recognition; image segmentation; and motion correction, registration, and tracking in space, time and/or spectral domains. We will help you visualize and capitalize on your images in ways you never thought possible.

Component Analysis

In many cases, clear signal-generation models do not exist or the model is too complicated to use. Our analysis methods extract hidden factors in your data waiting to be discovered. Our experience with principal, independent, nonnegative and other component-analysis approaches reveal which combinations of variables are needed to help you make the right decisions.

Stochastic Methods

In order to effectively deal with uncertainty, we employ a wide arsenal of probabilistic methods such as simulated annealing, stochastic optimization and multiple Monte Carlo approaches.

Machine Learning

Strong Predictive Analytics

Supervised

We work extensively with classification and prediction algorithms, applying them to complex data and also developing novel approaches and techniques.

Unsupervised

When outcome information is unavailable, we have the expert knowledge needed to choose an appropriate model for your data from large classes of clustering, manifold learning and mixture models.

Graphical Models

Casual information is commonly accessed via randomized experiments, while most of data are observational. We address the problem with models that can extract causal information from observational data.

Deep Models

Rarely do machine learning algorithms work well on complex raw data, relying on manual feature selection. We replace the manual process with the cutting edge automatic representation learning by deep models.

Statistics

Building confidence in your data by assessing its uncertainty

Power Analysis

How much data needs to be collected to achieve confidence in a given question? Is the sample large enough to support the claim? We will help you answer these and related questions.

Parametric Statistics

We are skilled in using an array of statistical tests and methods that will provide significant measures of the results based on your data.

Non-Parametric Statistics

Frequently, analytic descriptions of the processes governing your data are uncertain. In this case, non-parametric statistics apply. Relying on our experience in supercomputer-based (graphical processing unit) implementations, we will provide results efficiently and quickly.

Bayesian Analysis

We also have expertise in applying Bayesian approaches, which allow you to incorporate problem-specific information about how likely certain events are to occur, or to incorporate known constraints into the problems you are trying to solve.

Web and Database

Software Development

Software Development

Our specialists develop data analysis tools and applications in a range of high level programming languages and development environments including C++, Java, Python, Matlab, and R.

Web Development

For portability, we add web­based interfaces to our tools, making them accessible on any platform that can run a browser. We work with a range of standards and languages, including HTML,CSS, Javascript, D3, PHP, Ruby on Rail, SOAP and JSON.

Server Infrastructure

For deployment we setup and custom configure Apache web server and any required infrastructure using technologies such as LAMP/LAPP, Apache Tomcat and node.js.

Database Management

We base our data management and organization on PostgresSQL database, using our experience in SQL, Db4o and query based data mining.

Signal Processing

With decades of collective experience working with spatial and temporal data, our expertise in collecting and reconstructing data, signal compression, quality improvement via noise reduction, and extraction of important features will empower you to interpret and make predictions from your data.

Image Analysis

Our extensive experience handling images at all scales from 2D to 4D or higher-dimensional data includes object recognition; image segmentation; and motion correction, registration, and tracking in space, time and/or spectral domains. We will help you visualize and capitalize on your images in ways you never thought possible.

Component Analysis

In many cases, clear signal-generation models do not exist or the model is too complicated to use. Our analysis methods extract hidden factors in your data waiting to be discovered. Our experience with principal, independent, nonnegative and other component-analysis approaches reveal which combinations of variables are needed to help you make the right decisions.

Stochastic Methods

In order to effectively deal with uncertainty, we employ a wide arsenal of probabilistic methods such as simulated annealing, stochastic optimization and multiple Monte Carlo approaches.

Supervised

We work extensively with classification and prediction algorithms, applying them to complex data and also developing novel approaches and techniques.

Unsupervised

When outcome information is unavailable, we have the expert knowledge needed to choose an appropriate model for your data from large classes of clustering, manifold learning and mixture models.

Graphical Models

Casual information is commonly accessed via randomized experiments, while most of data are observational. We address the problem with models that can extract causal information from observational data.

Deep Models

Rarely do machine learning algorithms work well on complex raw data, relying on manual feature selection. We replace the manual process with the cutting edge automatic representation learning by deep models.

Power Analysis

How much data needs to be collected to achieve confidence in a given question? Is the sample large enough to support the claim? We will help you answer these and related questions.

Parametric Statistics

We are skilled in using an array of statistical tests and methods that will provide significant measures of the results based on your data.

Non-Parametric Statistics

Frequently, analytic descriptions of the processes governing your data are uncertain. In this case, non-parametric statistics apply. Relying on our experience in supercomputer-based (graphical processing unit) implementations, we will provide results efficiently and quickly.

Bayesian Analysis

We also have expertise in applying Bayesian approaches, which allow you to incorporate problem-specific information about how likely certain events are to occur, or to incorporate known constraints into the problems you are trying to solve.

Software Development

Our specialists develop data analysis tools and applications in a range of high level programming languages and development environments including C++, Java, Python, Matlab, and R.

Web Development

For portability, we add web­based interfaces to our tools, making them accessible on any platform that can run a browser. We work with a range of standards and languages, including HTML,CSS, Javascript, D3, PHP, Ruby on Rail, SOAP and JSON.

Server Infrastructure

For deployment we setup and custom configure Apache web server and any required infrastructure using technologies such as LAMP/LAPP, Apache Tomcat and node.js.

Database Management

We base our data management and organization on PostgresSQL database, using our experience in SQL, Db4o and query based data mining.