Development of Technology for Summarization of Kazakh Text

This paper presents the solution to problem of summarizing Kazakh texts. The text summarization is considered as a sequence two tasks: extracting most important sentences and simplifying received sentences. task solved using TF-IDF method neural network technology “Seq2Seq”. Problem NMT for simplification was in absence dataset training. To solve this work propose use transfer learning method. made it possible ready-made model that trained on parallel corpus Simple English Wikipedia not create from scratch. For this, language has been developed, based training language. Main scientific contribution simplification.