Machine learning of artificial intelligence agents on associative heterarchical memory

Natural language processing is an essential task for the whole field of artificial intelligence. To solve this problem, the authors proposed a new mathematical form, so called associative-heterarchical memory, based on the concept of hypergraph. Machine learning is a widely used method of artificial intelligence, especially when humans cannot determine patterns and frequencies of output values when viewing the data. Machine learning is best applicable for such tasks. Associative-heterarchical memory can also be applied to machine learning, and several methods (learning with a teacher, learning without a teacher, reinforcement learning) can be used to optimize the performance of associative-heterarchical memory-based artificial intelligence agents and perform required tasks. The output of such programs can be semantic or logical. Thus, machine learning is an important part of thewhole structure of associative-heterarchical memory. This article is devoted to interaction between associative-heterarchical memory and machine learning methods. Later on, the team of authors plans to write an additional article describing the output method using activation focus. This article will be of interest to experts in the area of artificial intelligence and mathematicians.