Prototyping Methodology of End-to-End Speech Analytics Software

Romanovskyi, O. and Iosifov, Ievgen and Iosifova, Olena and Sokolov, Volodymyr and Skladannyi, Pavlo and Sukaylo, Igor (2022) Prototyping Methodology of End-to-End Speech Analytics Software Modern Machine Learning Technologies and Data Science Workshop, 3312 (1). pp. 76-86. ISSN 1613-0073

[thumbnail of V_Sokolov_P_Skladannyi_I_Sukaylo_ta _in_V_MoMLeT&DS_2022_3312.pdf] Text
V_Sokolov_P_Skladannyi_I_Sukaylo_ta _in_V_MoMLeT&DS_2022_3312.pdf

Download (811kB)


This paper presents the prototype of end-to-end speech recognition, storage, and postprocessing tasks to build speech analytics, real-time agent augmentation, and other speechrelated products. Moving ASR models from the dev environment into production requires both researcher and architectural knowledge, which slows down and limits the possibility of companies benefiting from speech recognition and NLP advances for fundamental business operations. This paper proposes a fast and flexible prototype that can be easily implemented and used to serve ASR/NLP-trained models to solve business problems. Various software solutions’ compatibility problems were solved during the experimental setup assembly, and a working prototype was built and tested. An architectural diagram of the solution was also shown. Performance, limitations, and challenges of implementation are also described.

Item Type: Article
Uncontrolled Keywords: Natural Language Processing, NLP; Automatic Speech Recognition; ASR; speech analytics
Subjects: Це архівна тематика Київського університету імені Бориса Грінченка > Статті у наукометричних базах > Scopus
Divisions: Це архівні підрозділи Київського університету імені Бориса Грінченка > Факультет інформаційних технологій та математики > Кафедра інформаційної та кібернетичної безпеки імені професора Володимира Бурячка
Depositing User: Павло Миколайович Складанний
Date Deposited: 20 Feb 2023 07:46
Last Modified: 20 Feb 2023 07:46

Actions (login required)

View Item View Item