Votre recherche
Résultats 6 ressources
-
Anaouder mouezh e Brezhoneg gant Vosk
-
Whisper and other large-scale automatic speech recognition models have made significant progress in performance. However, their performance on many low-resource languages, such as Kazakh, is not satisfactory. It is worth researching how to utilize low-cost data to improve the performance of Whisper on under-represented languages. In this study, we utilized easily accessible unpaired speech and text data and combined the language model GPT with Whisper on Kazakh. We implemented end of transcript (EOT) judgment modification and hallucination penalty to improve the performance of speech recognition. Further, we employed the decoding average token log probability as a criterion to select samples from unlabeled speech data and used pseudo-labeled data to fine-tune the model to further improve its performance. Ultimately, we achieved more than 10\% absolute WER reduction in multiple experiments, and the whole process has the potential to be generalized to other under-represented languages.
-
This paper describes different approaches for developing, for the first time, an automatic speech recognition system for two of the main dialects of Occitan, namely Gascon and Languedocian, and the results obtained in them. The difficulty of the task lies in the fact that Occitan is a less-resourced language. Although a great effort has been made to collect or create corpora of each variant (transcribed speech recordings for the acoustic models and two text corpora for the language models), the sizes of the corpora obtained are far from those of successful systems reported in the literature, and thus we have tested different techniques to compensate for the lack of resources. We have developed classical systems using Kaldi, creating an acoustic model for each variant and also creating language models from the collected corpora and from machine translated texts. We have also tried fine-tuning a Whisper model with our speech corpora. We report word error rates of 20.86 for Gascon and 13.52 for Languedocian with the Kaldi systems and 16.37 for Gascon and 11.74 for Languedocian with Whisper.
-
We study the capabilities of speech processing systems trained simply to predict large amounts of transcripts of audio on the internet. When scaled to 680,000 hours of multilingual and multitask supervision, the resulting models generalize well to standard benchmarks and are often competitive with prior fully supervised results without the need for any dataset specific fine-tuning. When compared to humans, the models approach their accuracy and robustness. We are releasing models and inference code to serve as a foundation for further work on robust speech processing.
-
Identifying phone inventories is a crucial component in language documentation and the preservation of endangered languages. However, even the largest collection of phone inventory only covers about 2000 languages, which is only 1/4 of the total number of languages in the world. A majority of the remaining languages are endangered. In this work, we attempt to solve this problem by estimating the phone inventory for any language listed in Glottolog, which contains phylogenetic information regarding 8000 languages. In particular, we propose one probabilistic model and one non-probabilistic model, both using phylogenetic trees (“language family trees”) to measure the distance between languages. We show that our best model outperforms baseline models by 6.5 F1. Furthermore, we demonstrate that, with the proposed inventories, the phone recognition model can be customized for every language in the set, which improved the PER (phone error rate) in phone recognition by 25%.
-
Grapheme-to-Phoneme (G2P) has many applications in NLP and speech fields. Most existing work focuses heavily on languages with abundant training datasets, which limits the scope of target languages to less than 100 languages. This work attempts to apply zero-shot learning to approximate G2P models for all low-resource and endangered languages in Glottolog (about 8k languages). For any unseen target language, we first build the phylogenetic tree (i.e. language family tree) to identify top-k nearest languages for which we have training sets. Then we run models of those languages to obtain a hypothesis set, which we combine into a confusion network to propose a most likely hypothesis as an approximation to the target language. We test our approach on over 600 unseen languages and demonstrate it significantly outperforms baselines.