Detail publikace

Customization of Automatic Speech Recognition Engines for Rare Word Detection Without Costly Model Re-Training

BHATTACHARJEE, M. MOTLÍČEK, P. NIGMATULINA, I. HELMKE, H. OHNEISER, O. KLEINERT, M. EHR, H.

Originální název

Customization of Automatic Speech Recognition Engines for Rare Word Detection Without Costly Model Re-Training

Typ

článek ve sborníku mimo WoS a Scopus

Jazyk

angličtina

Originální abstrakt

Thanks to Alexa, Siri or Google Assistant automatic speech recognition (ASR) has changed our daily life during the last decade. Prototypic applications in the air traffic management (ATM) domain are available. Recently pre-filling radar label entries by ASR support has reached the technology readiness level before industrialization (TRL6). However, seldom spoken and airspace related words relevant in the ATM context remain a challenge for sophisticated applications. Open-source ASR toolkits or large pre-trained models for experts - allowing to tailor ASR to new domains - can be exploited with a typical constraint on availability of certain amount of domain specific training data, i.e., typically transcribed speech for adapting acoustic and/or language models. In general, it is sufficient for a "universal" ASR engine to reliably recognize a few hundred words that form the vocabulary of the voice communications between air traffic controllers and pilots. However, for each airport some hundred dependent words that are seldom spoken need to be integrated. These challenging word entities comprise special airline designators and waypoint names like "dexon" or "burok", which only appear in a specific region. When used, they are highly informative and thus require high recognition accuracies. Allowing plug and play customization with a minimum expert manipulation assumes that no additional training is required, i.e., fine-tuning the universal ASR. This paper presents an innovative approach to automatically integrate new specific word entities to the universal ASR system. The recognition rate of these region-specific word entities with respect to the universal ASR increases by a factor of 6.

Klíčová slova

Speech Recognition; Model Adaptation; Integration of prior knowledge; Customization of model, Rare-word integration.

Autoři

BHATTACHARJEE, M.; MOTLÍČEK, P.; NIGMATULINA, I.; HELMKE, H.; OHNEISER, O.; KLEINERT, M.; EHR, H.

Vydáno

27. 11. 2023

Nakladatel

SESAR Joint Undertaking

Místo

Seville

Strany od

1

Strany do

8

Strany počet

8

URL

BibTex

@inproceedings{BUT187995,
  author="BHATTACHARJEE, M. and MOTLÍČEK, P. and NIGMATULINA, I. and HELMKE, H. and OHNEISER, O. and KLEINERT, M. and EHR, H.",
  title="Customization of Automatic Speech Recognition Engines for Rare Word Detection Without Costly Model Re-Training",
  booktitle="Proceedings of the 13th SESAR Innovation Days",
  year="2023",
  pages="1--8",
  publisher="SESAR Joint Undertaking",
  address="Seville",
  doi="10.61009/SID.2023.1.10",
  url="https://www.sesarju.eu/sites/default/files/documents/sid/2023/Papers/SIDs_2023_paper_18%20final.pdf"
}