MEDC Company Profile

MEDC was created in 1981 by two Cambridge University academics: John Skilling, of the Department of Applied Mathematics and Theoretical Physics, and Steve Gull of the Department of Physics.

In the late 1970's, both had been involved in trying to process noisy radio-astronomical images using the newly discovered Maximum Entropy Method, or MEM. The application of MEM to a 1024-square astronomical image by Gull & Daniell in 1978 was a milestone, and showed that MEM was a practical technique, rather than a purely theoretical one.

It rapidly became obvious that MEM had much wider applicability than just image processing, and in fact, can generally be applied to the estimation of any Positive Additive Distribution (see our Products pages for many other examples of applications). MEDC was founded to exploit these other possibilities, and MEM rapidly found applications in fields as diverse as medical and seismic imaging,

The Maximum Entropy Method is based on Bayesian Estimation Theory, and is the natural choice of estimation technique when the number of variables (e.g pixel fluxes) is large. Nevertheless, there are still difficulties, especially in taking the formulation to the continuum limit. Massive Inference is a new technique which overcomes these using an atomic prior. For small numbers of variables MassInf yields superior results, but the running time is still often prohibitive for large problems, so that MEM then represents the optimum practical solution.

In 1990, MEDC spun-off another company, MaxEnt Solutions Limited (MSL). It was intended that MSL would handle the high-volume and OEM market, while MEDC retained the custom/bespoke side of the business. In 1997 MEDC absorbed Massive Inference Techniques (MIT), and in 1999, following several changes of personnel, MEDC re-acquired MSL. MEDC now handles the full range of business of all three companies.

Today, both MEDC principals continue to be deeply involved in the academic development of MEM and MassInf. Our code-base is regularly updated to reflect the latest developments in both the theory and numerical algorithms, and to take advantage of improvements in hardware.

As a company, we are committed to helping our customers achieve the best possible solution to their applications. We would like to hear from you.