---
title: "Arabic Dysarthric Speech Recognition Using Adversarial and Signal-Based Augmentation — Baali 2023"
description: "A post-processing step then replaces output words with Arabic Wikipedia lexicon entries, using a Jaccard distance weighted by… 18% WER. 34 references."
authors: ["Massa Baali", "Ibrahim Almakky", "Shady Shehata", "Fakhri Karray"]
year: 2023
venue: "arXiv:2306.04368"
source: "https://arxiv.org/abs/2306.04368"
doi: "10.48550/arxiv.2306.04368"
arxiv: "2306.04368"
licence: "CC BY-NC-SA 4.0 (arXiv)"
licence_url: "http://creativecommons.org/licenses/by-nc-sa/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/baali-2023-arabic-dysarthric-speech-recognition-using/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/baali-2023-arabic-dysarthric-speech-recognition-using/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2023](https://md.research.elamur.ai/free/en/papers/2023/page/7/) / Arabic Dysarthric Speech Recognition Using Adversarial and Signal-Based Augmentation

# Arabic Dysarthric Speech Recognition Using Adversarial and Signal-Based Augmentation

**Authors:** [Massa Baali](https://md.research.elamur.ai/free/en/authors/massa-baali/), Ibrahim Almakky, [Shady Shehata](https://md.research.elamur.ai/free/en/authors/shady-shehata/) and Fakhri Karray · **Year:** 2023 · **Venue:** arXiv:2306.04368 · **Licence:** [CC BY-NC-SA 4.0 (arXiv)](http://creativecommons.org/licenses/by-nc-sa/4.0/)

**Links:** [Publisher's page](https://arxiv.org/abs/2306.04368) · [PDF](https://arxiv.org/pdf/2306.04368) · [DOI 10.48550/arxiv.2306.04368](https://doi.org/10.48550/arxiv.2306.04368) · [arXiv 2306.04368](https://arxiv.org/abs/2306.04368) · [OpenAlex](https://openalex.org/W4385822655)

**Citations:** 4 citations (OpenAlex, 2026-10-14)

## Summary
Speech recognizers that work well for typical speakers fail on dysarthric speech, a motor disorder that makes articulation slow, slurred or prosodically irregular. In Arabic no dysarthric corpus exists, and recording patients is difficult because of fatigue and low compliance. The authors therefore synthesize training data. First, healthy Arabic recordings from Common Voice are altered with speed perturbation (resampling) and tempo perturbation (WSOLA). Second, a Parallel WaveGAN (PWG) trained on the English dysarthric Torgo corpus re-synthesizes these signals. This rests on the assumption that dysarthric traits are largely language-independent. The combined synthetic audio is used to fine-tune a Conformer previously trained on MGB-2. A post-processing step then replaces output words with Arabic Wikipedia lexicon entries, using a Jaccard distance weighted by character confusions. On synthetic dysarthric Common Voice speech the system reaches 18% WER and 17.2% CER, well ahead of a healthy-only baseline. A parallel English check on real dysarthric speech also shows large gains. The work offers a recipe for disordered-speech ASR when no in-language patient data exists.

## Key points
- Results: 18% WER and 17.2% CER on synthetically "dysarthrified" Arabic Common Voice, reported as an 81.8% WER improvement over a Conformer trained only on healthy speech.
- English validation: a baseline trained on healthy LJSpeech is compared with the augmented approach on real dysarthric speech, with a reported 124% WER improvement.
- Data: an 8-hour gender-balanced Common Voice Arabic subset (resampled to 16 kHz) is the augmentation source. A 20-hour MGB-2 subset (18 h train, 1 h dev, 1 h test) trains the base Arabic ASR. Torgo supplies English dysarthric speech, and a 4-hour LJSpeech subset is expanded into 16 hours of synthetic dysarthric English.
- Models: ESPnet Conformer (Conformer encoder plus Transformer decoder) for ASR. Parallel WaveGAN, with a WaveNet-style generator conditioned on mel-spectrograms plus noise, for adversarial augmentation.
- Text correction uses a weighted Jaccard distance in which frequently confused characters count as closer. Each predicted word is mapped to its nearest entry in a dictionary built from an Arabic Wikipedia dump.

## Interesting ideas
- Rare conditions can be simulated by applying simple, controllable distortions (here, speed and tempo) to plentiful clean data, giving a synthetic training set when real examples cannot be collected.
- The augmentation runs in two stages: a cheap rule-based transformation first, then a learned model that adds subtler properties the rules cannot express.
- The model's own observed confusions weight the edit distance used to snap noisy outputs onto a closed lexicon, so correction follows the errors the system actually makes.
- Patterns learned in a resource-rich language are transferred to a target language, on the assumption that the phenomenon is largely language-independent.
- The synthetic-data pipeline is validated in a language that has real impaired speech, as a sanity check before trusting purely synthetic evaluation in the target language.

## For Mova

How these ideas could serve Mova, our small open language model — its files, stages and tests:

- **I1** (tokens): Systematically corrupting clean treebank text with typos, dropped apostrophes or casing noise could give the tagger and lemmatizer training material for messy input, with each corruption type labeled and traceable.
- **I2** (learn): Silver data could be built the same way, with deterministic rewrite rules producing candidates and a learned or LLM teacher stage refining them, while keeping the rule layer inspectable.
- **I3** (tokens): A lemma or spelling normalizer could weight edits by the confusions logged in Mova's own tagging errors, and each correction would carry an explanation of the swap.
- **I4** (frames): Role or speech-act cues annotated richly in one resource could be tested for transfer to another genre or dialect, provided the transfer assumption is checked explicitly.
- **I5** (learn): Silver-data and repair-rule pipelines should be scored on a domain with gold UD annotation before results on silver-only domains are believed, much like a negative or positive control.

*Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.*

Fields: [Linguistics & languages](https://md.research.elamur.ai/free/en/shelves/linguistics/)

Topics: [🎙️ Speech & spoken language](https://md.research.elamur.ai/free/en/topics/speech/) · [🐪 Semitic languages](https://md.research.elamur.ai/free/en/topics/semitic-languages/)

## Cite

Cite as: Massa Baali, Ibrahim Almakky, Shady Shehata and Fakhri Karray. 2023. Arabic Dysarthric Speech Recognition Using Adversarial and Signal-Based Augmentation. arXiv:2306.04368.

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@misc{baali2023arabic,
  title = {{Arabic Dysarthric Speech Recognition Using Adversarial and Signal-Based Augmentation}},
  author = {Massa Baali, Ibrahim Almakky, Shady Shehata and Fakhri Karray},
  year = {2023},
  doi = {10.48550/arxiv.2306.04368},
  eprint = {2306.04368},
  archivePrefix = {arXiv},
  url = {https://doi.org/10.48550/arxiv.2306.04368}
}
```

- [← previous ALJP: An Arabic Legal Judgment Prediction in Personal Status Cases Using Machine Learning Models](https://md.research.elamur.ai/free/en/papers/abbara-2023-aljp-arabic-legal-judgment-prediction/)
- [next → Automatic Standardization of Arabic Dialects for Machine Translation](https://md.research.elamur.ai/free/en/papers/alnassan-2023-automatic-standardization-arabic-dialects-machine/)
- [↑ back to papers of 2023](https://md.research.elamur.ai/free/en/papers/2023/page/7/)

## References

34 references, 5 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. D. Amodei, S. Ananthanarayanan, R. Anubhai, J. Bai, E. Battenberg, C. Case, J. Casper, B. Catanzaro, Q. Cheng, G. Chen et al., “Deep speech 2: End-to-end speech recognition in english and mandarin,” in International conference on machine learning, 2016. [our card](https://md.research.elamur.ai/free/en/papers/amodei-2015-deep-speech-2-end-end/)
2. R. Turrisi, A. Braccia, M. Emanuele, S. Giulietti, M. Pugliatti, M. Sensi, L. Fadiga, and L. Badino, “Easycall corpus: a dysarthric speech dataset,” arXiv, 2021. [find](https://scholar.google.com/scholar?q=R%2E+Turrisi%2C+A%2E+Braccia%2C+M%2E+Emanuele%2C+S%2E+Giulietti%2C+M%2E+Pugliatti%2C+M%2E+Sensi%2C+L%2E+Fadiga%2C+and+L%2E+Badino%2C+%E2%80%9CEasycall+corpus%3A+a+dysarthric+speech+dataset%2C%E2%80%9D+arXiv%2C+2021%2E)
3. S. Hu, X. Xie, M. Geng, M. Cui, J. Deng, T. Wang, X. Liu, and H. Meng, “Exploiting cross-domain and cross-lingual ultrasound tongue imaging features for elderly and dysarthric speech recognition,” arXiv preprint arXiv:2206.07327, 2022. [↗](https://arxiv.org/abs/2206.07327)
4. Z. Yue, E. Loweimi, and Z. Cvetkovic, “Raw source and filter modelling for dysarthric speech recognition,” in ICASSP. IEEE, 2022. [find](https://scholar.google.com/scholar?q=Z%2E+Yue%2C+E%2E+Loweimi%2C+and+Z%2E+Cvetkovic%2C+%E2%80%9CRaw+source+and+filter+modelling+for+dysarthric+speech+recognition%2C%E2%80%9D+in+ICASSP%2E+IEEE%2C+2022%2E)
5. M. Geng, X. Xie, Z. Ye, T. Wang, G. Li, S. Hu, X. Liu, and H. Meng, “Speaker adaptation using spectro-temporal deep features for dysarthric and elderly speech recognition,” arXiv preprint arXiv:2202.10290, 2022. [↗](https://arxiv.org/abs/2202.10290)
6. N. M. Joy and S. Umesh, “Improving acoustic models in torgo dysarthric speech database,” IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2018. [find](https://scholar.google.com/scholar?q=N%2E+M%2E+Joy+and+S%2E+Umesh%2C+%E2%80%9CImproving+acoustic+models+in+torgo+dysarthric+speech+database%2C%E2%80%9D+IEEE+Transactions+on+Neural+Systems+and+Rehabilitation+Engineering%2C+2018%2E)
7. X. Xie, R. Ruzi, X. Liu, and L. Wang, “Variational auto-encoder based variability encoding for dysarthric speech recognition,” in Interspeech 2021. [find](https://scholar.google.com/scholar?q=X%2E+Xie%2C+R%2E+Ruzi%2C+X%2E+Liu%2C+and+L%2E+Wang%2C+%E2%80%9CVariational+auto%2Dencoder+based+variability+encoding+for+dysarthric+speech+recognition%2C%E2%80%9D+in+Interspeech+2021%2E)
8. A. Hernandez, P. A. Pérez-Toro, E. Nöth, J. R. Orozco-Arroyave, A. Maier, and S. H. Yang, “Cross-lingual self-supervised speech representations for improved dysarthric speech recognition,” in Interspeech 2022. [find](https://scholar.google.com/scholar?q=A%2E+Hernandez%2C+P%2E+A%2E+Pe%CC%81rez%2DToro%2C+E%2E+No%CC%88th%2C+J%2E+R%2E+Orozco%2DArroyave%2C+A%2E+Maier%2C+and+S%2E+H%2E+Yang%2C+%E2%80%9CCross%2Dlingual+self%2Dsupervised+speech+representations+for+improved+dysarthric+speech+recognition%2C%E2%80%9D+in+Inters)
9. C. Espana-Bonet and J. A. Fonollosa, “Automatic speech recognition with deep neural networks for impaired speech,” in International Conference on Advances in Speech and Language Technologies for Iberian Languages. Springer, 2016. [find](https://scholar.google.com/scholar?q=C%2E+Espana%2DBonet+and+J%2E+A%2E+Fonollosa%2C+%E2%80%9CAutomatic+speech+recognition+with+deep+neural+networks+for+impaired+speech%2C%E2%80%9D+in+International+Conference+on+Advances+in+Speech+and+Language+Technologies+for+Iberi)
10. M. Soleymanpour, M. T. Johnson, R. Soleymanpour, and J. Berry, “Synthesizing dysarthric speech using multi-speaker tts for dysarthric speech recognition,” in ICASSP, 2022. [find](https://scholar.google.com/scholar?q=M%2E+Soleymanpour%2C+M%2E+T%2E+Johnson%2C+R%2E+Soleymanpour%2C+and+J%2E+Berry%2C+%E2%80%9CSynthesizing+dysarthric+speech+using+multi%2Dspeaker+tts+for+dysarthric+speech+recognition%2C%E2%80%9D+in+ICASSP%2C+2022%2E)
11. J. Ren and M. Liu, “An automatic dysarthric speech recognition approach using deep neural networks,” International Journal of Advanced Computer Science and Applications, 2017. [find](https://scholar.google.com/scholar?q=J%2E+Ren+and+M%2E+Liu%2C+%E2%80%9CAn+automatic+dysarthric+speech+recognition+approach+using+deep+neural+networks%2C%E2%80%9D+International+Journal+of+Advanced+Computer+Science+and+Applications%2C+2017%2E)
12. Y. Matsuzaka, R. Takashima, C. Sasaki, and T. Takiguchi, “Data augmentation for dysarthric speech recognition based on text-to-speech synthesis,” in 2022 IEEE 4th Global Conference on Life Sciences and Technologies. [find](https://scholar.google.com/scholar?q=Y%2E+Matsuzaka%2C+R%2E+Takashima%2C+C%2E+Sasaki%2C+and+T%2E+Takiguchi%2C+%E2%80%9CData+augmentation+for+dysarthric+speech+recognition+based+on+text%2Dto%2Dspeech+synthesis%2C%E2%80%9D+in+2022+IEEE+4th+Global+Conference+on+Life+Sciences+an)
13. M. Dhanalakshmi and P. Vijayalakshmi, “Intelligibility modification of dysarthric speech using hmm-based adaptive synthesis system,” in ICoBE. IEEE, 2015. [find](https://scholar.google.com/scholar?q=M%2E+Dhanalakshmi+and+P%2E+Vijayalakshmi%2C+%E2%80%9CIntelligibility+modification+of+dysarthric+speech+using+hmm%2Dbased+adaptive+synthesis+system%2C%E2%80%9D+in+ICoBE%2E+IEEE%2C+2015%2E)
14. B. Vachhani, C. Bhat, and S. K. Kopparapu, “Data augmentation using healthy speech for dysarthric speech recognition.” in Interspeech, 2018. [find](https://scholar.google.com/scholar?q=B%2E+Vachhani%2C+C%2E+Bhat%2C+and+S%2E+K%2E+Kopparapu%2C+%E2%80%9CData+augmentation+using+healthy+speech+for+dysarthric+speech+recognition%2E%E2%80%9D+in+Interspeech%2C+2018%2E)
15. M. Geng, X. Xie, S. Liu, J. Yu, S. Hu, X. Liu, and H. Meng, “Investigation of data augmentation techniques for disordered speech recognition,” Interspeech, pp. 696–700, 2020. [find](https://scholar.google.com/scholar?q=M%2E+Geng%2C+X%2E+Xie%2C+S%2E+Liu%2C+J%2E+Yu%2C+S%2E+Hu%2C+X%2E+Liu%2C+and+H%2E+Meng%2C+%E2%80%9CInvestigation+of+data+augmentation+techniques+for+disordered+speech+recognition%2C%E2%80%9D+Interspeech%2C+pp%2E+696%E2%80%93700%2C+2020%2E)
16. C. Bhat, A. Panda, and H. Strik, “Improved asr performance for dysarthric speech using two-stage dataaugmentation,” Interspeech 2022. [find](https://scholar.google.com/scholar?q=C%2E+Bhat%2C+A%2E+Panda%2C+and+H%2E+Strik%2C+%E2%80%9CImproved+asr+performance+for+dysarthric+speech+using+two%2Dstage+dataaugmentation%2C%E2%80%9D+Interspeech+2022%2E)
17. Z. Jin, M. Geng, X. Xie, J. Yu, S. Liu, X. Liu, and H. Meng, “Adversarial data augmentation for disordered speech recognition,” Proc. Interspeech 2021. [find](https://scholar.google.com/scholar?q=Z%2E+Jin%2C+M%2E+Geng%2C+X%2E+Xie%2C+J%2E+Yu%2C+S%2E+Liu%2C+X%2E+Liu%2C+and+H%2E+Meng%2C+%E2%80%9CAdversarial+data+augmentation+for+disordered+speech+recognition%2C%E2%80%9D+Proc%2E+Interspeech+2021%2E)
18. J. Yu, X. Xie, S. Liu, S. Hu, M. W. Lam, X. Wu, K. H. Wong, X. Liu, and H. Meng, “Development of the cuhk dysarthric speech recognition system for the ua speech corpus.” in Interspeech, 2018. [find](https://scholar.google.com/scholar?q=J%2E+Yu%2C+X%2E+Xie%2C+S%2E+Liu%2C+S%2E+Hu%2C+M%2E+W%2E+Lam%2C+X%2E+Wu%2C+K%2E+H%2E+Wong%2C+X%2E+Liu%2C+and+H%2E+Meng%2C+%E2%80%9CDevelopment+of+the+cuhk+dysarthric+speech+recognition+system+for+the+ua+speech+corpus%2E%E2%80%9D+in+Interspeech%2C+2018%2E)
19. H. Kim, M. Hasegawa-Johnson, A. Perlman, J. Gunderson, T. S. Huang, K. Watkin, and S. Frame, “Dysarthric speech database for universal access research,” in Ninth Annual Conference of the International Speech Communication Association, 2008. [find](https://scholar.google.com/scholar?q=H%2E+Kim%2C+M%2E+Hasegawa%2DJohnson%2C+A%2E+Perlman%2C+J%2E+Gunderson%2C+T%2E+S%2E+Huang%2C+K%2E+Watkin%2C+and+S%2E+Frame%2C+%E2%80%9CDysarthric+speech+database+for+universal+access+research%2C%E2%80%9D+in+Ninth+Annual+Conference+of+the+International)
20. D. Wang, J. Yu, X. Wu, L. Sun, X. Liu, and H. Meng, “Improved end-to-end dysarthric speech recognition via meta-learning based model re-initialization,” in ISCSLP. IEEE, 2021. [find](https://scholar.google.com/scholar?q=D%2E+Wang%2C+J%2E+Yu%2C+X%2E+Wu%2C+L%2E+Sun%2C+X%2E+Liu%2C+and+H%2E+Meng%2C+%E2%80%9CImproved+end%2Dto%2Dend+dysarthric+speech+recognition+via+meta%2Dlearning+based+model+re%2Dinitialization%2C%E2%80%9D+in+ISCSLP%2E+IEEE%2C+2021%2E)
21. R. Yamamoto, E. Song, and J.-M. Kim, “Parallel wavegan: A fast waveform generation model based on generative adversarial networks with multi-resolution spectrogram,” in ICASSP. IEEE, 2020. [find](https://scholar.google.com/scholar?q=R%2E+Yamamoto%2C+E%2E+Song%2C+and+J%2E%2DM%2E+Kim%2C+%E2%80%9CParallel+wavegan%3A+A+fast+waveform+generation+model+based+on+generative+adversarial+networks+with+multi%2Dresolution+spectrogram%2C%E2%80%9D+in+ICASSP%2E+IEEE%2C+2020%2E)
22. M. A. Hegazi and A. Abdelhamid, “Developing an arabic speech intelligibility test for adolescents and adults,” The Egyptian Journal of Otolaryngology, vol. 35, no. 1, pp. 86–91, 2019. [find](https://scholar.google.com/scholar?q=M%2E+A%2E+Hegazi+and+A%2E+Abdelhamid%2C+%E2%80%9CDeveloping+an+arabic+speech+intelligibility+test+for+adolescents+and+adults%2C%E2%80%9D+The+Egyptian+Journal+of+Otolaryngology%2C+vol%2E+35%2C+no%2E+1%2C+pp%2E+86%E2%80%9391%2C+2019%2E)
23. T. Ko, V. Peddinti, D. Povey, and S. Khudanpur, “Audio augmentation for speech recognition,” in Sixteenth annual conference of the international speech communication association, 2015. [find](https://scholar.google.com/scholar?q=T%2E+Ko%2C+V%2E+Peddinti%2C+D%2E+Povey%2C+and+S%2E+Khudanpur%2C+%E2%80%9CAudio+augmentation+for+speech+recognition%2C%E2%80%9D+in+Sixteenth+annual+conference+of+the+international+speech+communication+association%2C+2015%2E)
24. W. Verhelst and M. Roelands, “An overlap-add technique based on waveform similarity (wsola) for high quality time-scale modification of speech,” in 1993 IEEE International Conference on Acoustics, Speech, and Signal Processing. [find](https://scholar.google.com/scholar?q=W%2E+Verhelst+and+M%2E+Roelands%2C+%E2%80%9CAn+overlap%2Dadd+technique+based+on+waveform+similarity+%28wsola%29+for+high+quality+time%2Dscale+modification+of+speech%2C%E2%80%9D+in+1993+IEEE+International+Conference+on+Acoustics%2C+Spe)
25. P. Guo, F. Boyer, X. Chang, T. Hayashi, Y. Higuchi, H. Inaguma, N. Kamo, C. Li, D. Garcia-Romero, J. Shi et al., “Recent developments on espnet toolkit boosted by conformer,” in ICASSP 2021. [find](https://scholar.google.com/scholar?q=P%2E+Guo%2C+F%2E+Boyer%2C+X%2E+Chang%2C+T%2E+Hayashi%2C+Y%2E+Higuchi%2C+H%2E+Inaguma%2C+N%2E+Kamo%2C+C%2E+Li%2C+D%2E+Garcia%2DRomero%2C+J%2E+Shi+et+al%2E%2C+%E2%80%9CRecent+developments+on+espnet+toolkit+boosted+by+conformer%2C%E2%80%9D+in+ICASSP+2021%2E)
26. J. M. Hancock, “Jaccard distance (jaccard index, jaccard similarity coefficient),” Dictionary of Bioinformatics and Computational Biology, 2004. [find](https://scholar.google.com/scholar?q=J%2E+M%2E+Hancock%2C+%E2%80%9CJaccard+distance+%28jaccard+index%2C+jaccard+similarity+coefficient%29%2C%E2%80%9D+Dictionary+of+Bioinformatics+and+Computational+Biology%2C+2004%2E)
27. A. Khooli, “Arabic Wiki data Dump 2018 Kaggle,” [Online; accessed 22-Oct-2022]. [Online]. Available: [https://www.kaggle.com/datasets/abedkhooli/arabic-wiki-data-dump-2018](https://www.kaggle.com/datasets/abedkhooli/arabic-wiki-data-dump-2018) [↗](https://www.kaggle.com/datasets/abedkhooli/arabic-wiki-data-dump-2018)
28. R. Ardila, M. Branson, K. Davis, M. Henretty, M. Kohler, J. Meyer, R. Morais, L. Saunders, F. M. Tyers, and G. Weber, “Common voice: A massively-multilingual speech corpus,” in Proceedings of the 12th Language Resources and Evaluation Conference, 2020. [our card](https://md.research.elamur.ai/free/en/papers/ardila-2019-common-voice-massively-multilingual-speech/)
29. A. Ali, P. Bell, J. Glass, Y. Messaoui, H. Mubarak, S. Renals, and Y. Zhang, “The mgb-2 challenge: Arabic multi-dialect broadcast media recognition,” in SLT 2016. [our card](https://md.research.elamur.ai/free/en/papers/ali-2016-mgb-2-challenge-arabic-multi/)
30. F. Rudzicz, A. K. Namasivayam, and T. Wolff, “The torgo database of acoustic and articulatory speech from speakers with dysarthria,” Language Resources and Evaluation, 2012. [find](https://scholar.google.com/scholar?q=F%2E+Rudzicz%2C+A%2E+K%2E+Namasivayam%2C+and+T%2E+Wolff%2C+%E2%80%9CThe+torgo+database+of+acoustic+and+articulatory+speech+from+speakers+with+dysarthria%2C%E2%80%9D+Language+Resources+and+Evaluation%2C+2012%2E)
31. K. Ito and L. Johnson, “The lj speech dataset,” Online: [https://keithito.com/LJ-Speech-Dataset](https://keithito.com/LJ-Speech-Dataset), 2017. [↗](https://keithito.com/LJ-Speech-Dataset)
32. T. Kudo and J. Richardson, “Sentencepiece: A simple and language independent subword tokenizer and detokenizer for neural text processing,” EMNLP, p. 66, 2018. [our card](https://md.research.elamur.ai/free/en/papers/kudo-2018-sentencepiece-simple-language-independent-subword/)
33. A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, L. Kaiser, and I. Polosukhin, “Attention is all you need,” Advances in neural information processing systems, 2017. [our card](https://md.research.elamur.ai/free/en/papers/vaswani-2017-attention-all-need/)
34. C. Lo, S. Fu, W. Huang, X. Wang, J. Yamagishi, Y. Tsao, and H. Wang, “Mosnet: Deep learning based objective assessment for voice conversion,” in Interspeech 2019. [find](https://scholar.google.com/scholar?q=C%2E+Lo%2C+S%2E+Fu%2C+W%2E+Huang%2C+X%2E+Wang%2C+J%2E+Yamagishi%2C+Y%2E+Tsao%2C+and+H%2E+Wang%2C+%E2%80%9CMosnet%3A+Deep+learning+based+objective+assessment+for+voice+conversion%2C%E2%80%9D+in+Interspeech+2019%2E)