Deep Transformer based Data Augmentation with Subword Units for Morphologically Rich Online ASR

B Tarján, G Szaszák, T Fegyó, P Mihajlik - arXiv preprint arXiv:2007.06949, 2020 - arxiv.org
Recently Deep Transformer models have proven to be particularly powerful in language
modeling tasks for ASR. Their high complexity, however, makes them very difficult to apply in …

Morphology aware data augmentation with neural language models for online hybrid ASR

B Tarján, T Fegyó, P Mihajlik - Acta Linguistica Academica, 2022 - akjournals.com
Recognition of Hungarian conversational telephone speech is challenging due to the
informal style and morphological richness of the language. Neural Network Language …

Improving real-time recognition of morphologically rich speech with transformer language model

B Tarján, G Szaszák, T Fegyó… - 2020 11th IEEE …, 2020 - ieeexplore.ieee.org
Transformer models have become to state-of-the-art in natural language understanding,
their use for language modeling in Automatic Speech Recognition (ASR) is also promising …

On the effectiveness of neural text generation based data augmentation for recognition of morphologically rich speech

B Tarján, G Szaszák, T Fegyó, P Mihajlik - Text, Speech, and Dialogue …, 2020 - Springer
Advanced neural network models have penetrated Automatic Speech Recognition (ASR) in
recent years, however, in language modeling many systems still rely on traditional Back-off …

Language Modeling for Hungarian Speech Recognition

B Tarján - 2021 - search.proquest.com
Automatic speech recognition (ASR) systems enable the machine transcription of human
speech. Language model plays an important role in ASR systems, as the final automatic …