Principles of compression and decompression within the form of multidimensional language representation applied to educational tasks generation

In the given work the problem of semantic decompression patterns addition to the meaningful sentences generated as functions in multidimensional vector space of the notions of the natural language is considered. The model of adding semantic decompression patterns based on stylistically-oriented sets of generative grammar rules is offered. The complexity of natural language translation problem is discussed. The considered model provides the algorithm of adding semantic decompression patterns that can be used for improving the performance of educational tasks generation software.