Intelligent support development of educational programs based on the neural language models taking into account of the labor market requirements

The active development of the digital economy today imposes high requirements on the adaptability, practical orientation and quality of educational content. Existing approaches to the intelligent decision support of the formation of educational programs based on ontological models, expert systems and heuristic algorithms do not allow effectively taking into account and tracking changes both in the labor market and in the open educational content space in the Massive Open Online Courses (MOOC). Instead, it is proposed to use approaches to the semantic analysis based on the well-known neural network language model word2vec, which is trained without supervision on large text corpora. The complexity of semantic analysis is to move the definition of semantic similarity measures for short texts of the extracted entities (course topics, learning outcomes, job requirements, etc.) to matching of large structured documents, such as professional standard, an educational program. To take into account the interrelations of entities, a graph model is introduced for representing the educational and professional domain. The paper proposes an artificial intelligent method of forming recommendations for the actualization of the learning outcomes and content of educational programs. At the first stage, the actual requirements of the labor market are determined based on a semantic matching of job requirements with the content of professional standards. The second stage includes a semantic matching of the content of academic disciplines with the requirements of the labor market. At the third stage, a semantic search of relevant educational content is carried out among the programs of disciplines of leading universities and massive open online courses (MOOC). During the fourth stage, final recommendations on updating the educational program are formed. The experiment demonstrated the possibility of applying the method for matching learning outcomes and content of disciplines with the requirements of professional standards and evaluation using the example of the educational program (bachelor degree) of computer science and engineering.