Continuous Growth of Information and the “Forgetting” Function in the Recognition of Sound Information

The paper presents the results of experiments with the most stable results in such a crucial issue as teaching the system to new visual and audio objects of information. In parallel, a mechanism is presented that allows the system to move to a new level of learning - self-learning. The developed self-learning method is a delayed learning mechanism, when the system learns to recognize new objects without a special command and when information characterizing a new object can enter the system as needed (depending on the system operator). The next stage in the development of the system was the need to solve the problem of the rapid growth of information volumes and, accordingly, the shortage of disk capacities. The “forgetting” function being developed has three or four ways of implementation and the problems being solved allow us to look from the inside (drawing analogies from the point of view of software) at how human intelligence resorts to solve this problem to build similar abilities. At the end of the work, classes of visual objects are generally studied, which have not yet received a sustainable solution, and private methods need artificial adjustments with subsequent difficulties with integration into the general and quite harmonious mode of operation of the system.