The problems I have solved on a scientific level do not directly translate into software solutions. In fact, a patent lawyer told me at the time, I can’t patent a theory but must create ‘vendible products’. Subsequently I submitted five patent applications, but was advised to withdraw them, as “patents are the game of the big boys”. As web editor of the SME Innovation Alliance, I have met the inventor who burnt his patents outside Westminster to demonstrate its usefulness!…
However, here are the problems that my software methods solve:
- Forecasting Methods are neither particularly accurate nor are they so universal that they can be applied to any application and any time interval. Using the same algorithm for a variety of ‘use cases’ is obviously highly advantageous in terms of code development.
- Multi-dimensional Data can, so far, be visualized in many ‘sexy’ ways but not such that mathematical correlations can be visualized. The unique combination of visual patterns and numerical quantifications, enabled by software parameters opens a whole new gamut of possibilities for exploration and investigation – into a lot of old and new data out there in all those clouds.
- Images can’t be quantified, I was told. But my algorithms make it possibly that they can be
- classified, sorted and ranked
- selected according to selection criteria
- analysed according to shapes, objects, patterns and other characteristics and thus compared – both individually and in series.
This added functionality is a solution to an image management problem that allows for finding hairline fractures in thousands of x-rays just as much as the monitoring of therapeutic processes in rare diseases:
- automated or ‘high throughput’ image processing becomes more precise – through a new kind of expert system that I call ‘smart knowledge portal’ – since human knowledge is combined with smart software and the strength of computers: number crunching.