Abstract:
The relevance of the stated subject of this scientific research is determined by the numerous
difficulties of spontaneous speech recognition due to the presence of a complex of unrelated
factors, as well as the need to find the best ways to overcome them through recognition
programs that are used in the development of a multilingual corpus. This research aims to
identify and study the key features of spontaneous speech that should be considered when
creating recognition programs. The basis of the methodological approach in this scientific
study is a combination of methods combining a systematic analysis of the principles of
building a multilingual corpus with an analytical study of the principles of operation of
spontaneous speech recognition programs. In the course of this scientific study, results were
obtained that indicate a significant increase in the practical efficiency of automatic speech
recognition systems when introducing sound signal correction algorithms, which opens up
additional opportunities for developing such software. In addition, the results of this
scientific study clearly demonstrate the significant impact of the accuracy of the effects of
speech recognition programs on the quality of development of multilingual corpora, which
include relatively large volumes of texts. The practical significance of the results obtained
in the course of this scientific work, as well as the conclusions formulated on their basis,
lies in the possibility of their use in the development of a multilingual corpus and
spontaneous speech recognition programs for their subsequent use in various information
systems, to obtain results related to the need accurate decoding of automatic speech, taking
into account all its characteristic features.
