10171045Generic AI: AGI - MVPClosedFast Start ResponseInnovate UKWe are creating a prototype AGI system. The AGI will use a new form of A.I. that we call _Generic AI_. _Generic AI_ works by applying a minimum-principle in Information and is constructed using Set theory. In order to form an AGI the system must _'think for itself'_. Therefore it is important that the system has no predetermined rules or functions. _Generic AI_ uses the same method to analyse _ANY_ kind of input. Consequently, all the patterns and sequences that it extracts from any kind of data-input can be compared against patterns and sequences from any other kind of data input. Unlike neural networks, the measurements and analysis that _Generic AI_ derives from data are Universal and Objective. By contrast neural networks use arbitrary mathematical functions expressed using calculus. This restricts the way that neural networks operate and predetermines the outcome - thereby making it impossible for the system to act as an independent '_intelligence_' and to _'think for itself'._ The AGI will accept any kind of input then will _Project_ that input ahead a number of times with random variations. The system will have a database that can be set up ahead of this Projection forming a '_Prior_' Set. This '_Prior_' Set will influence the projected outcomes by including this data. The more detail that is included in this '_Prior_' Set, the more accurate the _Projections_ will be. The aim of the system is to find the minimum pathway ahead in Union with the _Prior Set._ This same algorithm will work for ANY kind of scenario. For example, if the Prior set is an image database and the inputted '_Problem_' is an image then this process will find the best matching image. The AGI algorithm is modelled on a profound theory of information which represents '_Problems_' and '_Solutions_' in terms of the Information in the set describing the inputted data. This is based on the concept that a '_Solution_' is the Set which, - including all the details of the '_Problem_' Set, - has smallest Cardinality. However, the AGI system entirely depends on the method of data analysis being 'Objective' and having 'Universal' values. This would not be possible with neural networks but it is possible with the _Generic AI_ system.