The following is a sampling of the types of solutions we have created for our clients:1
Machine Learning Frameworks
We developed a set of end-to-end, pipeline frameworks to help our clients design, train and test machine learning algorithms. The pipeline frameworks were tailored to each client’s specific needs and custom user interfaces were created for processing data, annotating data and evaluating the resulting algorithms.
We created MATLAB-based user interfaces to aid in the collection of experimental data in a product test situation. The user interfaces support both calibration and production runs and interfaced with several measuring devices. The data was displayed in graphical and tabular form and saved in a database. The GUI was designed to be run by technicians with little or no MATLAB experience.
M-file Speed Improvement
MATLAB m-file speed-up is one of our most popular services. We routinely make speed improvements of 2x or more. For one particular client, we made changes to several m-files used within a long running Monte-Carlo simulation and achieved a substantial speed improvement. A simulation that used to take a day could now be done in minutes. The speed improvement allowed the client to run many more runs within the allotted time and to avoid purchasing additional workstations.
We created a MATLAB-based user interface and m-files as a front end to a set of m-files developed by university researchers. The GUI simplifies the specification of the research parameters and keeps track of the various runs. The results are plotted within an interactive graph GUI that allows the researchers to quickly view various aspects of the results.
Custom Numerical Algorithm
We developed a custom numerical algorithm to solve a physics-based mathematical problem that the standard algorithms failed to solve. The resulting algorithm was able to handle various singularities and non-linear behavior.
We helped a researcher turn a mathematical model into a numerical model in MATLAB. The model was based on ordinary and partial differential equations. The researcher subsequently used the model to compare to experimental results taken from actual hardware and publish the results.
We have helped several clients using MATLAB to find optimal solutions. Our advice helped the clients improve convergence and find better solutions. In some cases, we suggested alternative optimization techniques that allowed them to find a better solution or scale up to handle a more complicated problem.
Custom M-file Development
Many clients have asked us to write m-files to read proprietary data formats, perform data analysis and visualization and implement an algorithms in MATLAB.
To inquire about our services without obligation, please contact Dr. Clay Thompson using our Contact Form.
1The case studies on this page have been generalized and obscured to protect the confidentiality of our clients