Inserting Punctuation to ASR Output in a Real-Time Production Environment

Result code in IS VaVaI
<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F28479777%3A_____%2F20%3AN0000001" target="_blank" >RIV/28479777:_____/20:N0000001 - isvavai.cz</a>
Result on the web
<a href="https://link.springer.com/chapter/10.1007/978-3-030-58323-1_45" target="_blank" >https://link.springer.com/chapter/10.1007/978-3-030-58323-1_45</a>
DOI - Digital Object Identifier
<a href="http://dx.doi.org/10.1007/978-3-030-58323-1_45" target="_blank" >10.1007/978-3-030-58323-1_45</a>

Result language
angličtina
Original language name
Inserting Punctuation to ASR Output in a Real-Time Production Environment
Original language description
The output of a speech recognition system is a continuous stream of words that has to be post-processed in various ways, out of which punctuation insertion is an essential step. Punctuated text is far more comprehensible to the reader, can be used for subtitling, and is necessary for further NLP processing, such as machine translation. In this article, we describe how state-of-the-art results in the field of punctuation restoration can be utilized in a production-ready business environment in the Czech language. A recurrent neural network based on long short-term memory is employed, making use of various features: textual based on pre-trained word embeddings, prosodic (mainly temporal), morphological, noise information, and speaker diarization. All the features except morphological tags were found to improve our baseline system. As we work in a real-time setup, it is not possible to employ information from the future of the word stream, yet we achieve significant improvements using LSTM. The usage of RNN also allows the model to learn longer dependencies than any n-gram-based language model can, which we find essential for insertion of question marks. The deployment of an RNN-based model thus leads to a relative 22.6 % decrease in punctuation errors and improvement in all metrics but one.
Czech name
—
Czech description
—

Type
D - Article in proceedings
CEP classification
—
OECD FORD branch
10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Project
<a href="/en/project/FW01010468" target="_blank" >FW01010468: Beey Multimedia Platform</a><br>
Continuities
P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)

Publication year
2020
Confidentiality
C - Předmět řešení projektu podléhá obchodnímu tajemství (§ 504 Občanského zákoníku), ale název projektu, cíle projektu a u ukončeného nebo zastaveného projektu zhodnocení výsledku řešení projektu (údaje P03, P04, P15, P19, P29, PN8) dodané do CEP, jsou upraveny tak, aby byly zveřejnitelné.

Similar results(10)