Application of artificial intelligence in search in patent and technical literature

A new direction of research in patent search is considered-automatic search for" similar " patent documents, search for documents that characterize the state of the art in a specific subject area, in the tasks of examination of inventions and utility models. A new quality criterion for patent search for examination purposes is described. The first results of the search for similar ones using artificial intelligence methods and tools obtained by the authors for searching through arrays of patent documents in Russian are presented. The directions of development of the used methods for search in English-language arrays, and in the future for multilingual search, as well as for some search scenarios in scientific and technical literature are proposed.