mlsanityData to 7 Oct 2026

ASR2K: Speech Recognition for Around 2000 Languages without Audio

arXiv 2209.02842 ↗First posted 6 Sep 2022cs.CLcited by 26

Not on the map. Why?

Abstract (arXiv)

Most recent speech recognition models rely on large supervised datasets, which are unavailable for many low-resource languages. In this work, we present a speech recognition pipeline that does not require any audio for the target language. The only assumption is that we have access to raw text datasets or a set of n-gram statistics. Our speech pipeline consists of three components: acoustic, pronunciation, and language models. Unlike the standard pipeline, our acoustic and pronunciation models use multilingual models without any supervision. The language model is built using n-gram statistics or the raw text dataset. We build speech recognition for 1909 languages by combining it with Crubadan: a large endangered languages n-gram database. Furthermore, we test our approach on 129 languages across two datasets: Common Voice and CMU Wilderness dataset. We achieve 50% CER and 74% WER on the Wilderness dataset with Crubadan statistics only and improve them to 45% CER and 69% WER when using 10000 raw text utterances.

Key facts. Cited by 26 papers in the map (data through 7 Oct 2026). Builds on 23 of its 34 references that are in the map. First posted on arXiv 6 Sep 2022.

Builds on

23 of 34 references.

References · 23

  1. 2020wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representationscited by 3,853
  2. 2014Sequence to Sequence Learning with Neural Networkscited by 2,974
  3. 2021HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Unitscited by 2,301
  4. 2020Conformer: Convolution-augmented Transformer for Speech Recognitioncited by 1,838
  5. 2019Common Voice: A Massively-Multilingual Speech Corpuscited by 1,358
  6. 2013Speech Recognition with Deep Recurrent Neural Networkscited by 666
  7. 2021SUPERB: Speech processing Universal PERformance Benchmarkcited by 652
  8. 2018ESPnet: End-to-End Speech Processing Toolkitcited by 554
  9. 2020Unsupervised Cross-lingual Representation Learning for Speech Recognitioncited by 496
  10. 2017Word Translation Without Parallel Datacited by 480
  11. 2019A Comparative Study on Transformer vs RNN in Speech Applicationscited by 210
  12. 2021Unsupervised Speech Recognitioncited by 147
11 more
  1. 2018A robust self-learning method for fully unsupervised cross-lingual mappings of word embeddingscited by 145
  2. 2019End-to-end ASR: from Supervised to Semi-Supervised Learning with Modern Architecturescited by 120
  3. 2019Unsupervised speech representation learning using WaveNet autoencoderscited by 90
  4. 2020Universal Phone Recognition with a Multilingual Allophone Systemcited by 81
  5. 2015EESEN: End-to-End Speech Recognition using Deep RNN Models and WFST-based Decodingcited by 73
  6. 2019Speech Recognition with Augmented Synthesized Speechcited by 60
  7. 2016Wav2Letter: an End-to-End ConvNet-based Speech Recognition Systemcited by 43
  8. 2020LRSpeech: Extremely Low-Resource Speech Synthesis and Recognitioncited by 43
  9. 2019VQVAE Unsupervised Unit Discovery and Multi-scale Code2Spec Inverter for Zerospeech Challenge 2019cited by 25
  10. 2020AlloVera: A Multilingual Allophone Databasecited by 10
  11. 2019Multilingual Speech Recognition with Corpus Relatedness Samplingcited by 3

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2022: 3; 2023: 9; 2024: 7; 2025: 5; 2026: 2.

26 papers.

Sort:

  1. 2023Scaling Speech Technology to 1,000+ Languagescited by 401
  2. 2023Google USM: Scaling Automatic Speech Recognition Beyond 100 Languagescited by 228
  3. 2023ML-SUPERB: Multilingual Speech Universal PERformance Benchmarkcited by 70
  4. 2024Towards Robust Speech Representation Learning for Thousands of Languagescited by 48
  5. 2023Improving Massively Multilingual ASR With Auxiliary CTC Objectivescited by 31
  6. 2024ML-SUPERB 2.0: Benchmarking Multilingual Speech Models Across Modeling Constraints, Languages, and Datasetscited by 26
  7. 2022Textless Direct Speech-to-Speech Translation with Discrete Speech Representationcited by 19
  8. 2022Towards Zero-Shot Code-Switched Speech Recognitioncited by 15
  9. 2025Charting the Landscape of African NLP: Mapping Progress and Shaping the Road Aheadcited by 15
  10. 2024Anatomy of Industrial Scale Multilingual ASRcited by 14
  11. 2023Allophant: Cross-lingual Phoneme Recognition with Articulatory Attributescited by 13
  12. 2025CS-FLEURS: A Massively Multilingual and Code-Switched Speech Datasetcited by 11
14 more
  1. 2023Learning to Speak from Text: Zero-Shot Multilingual Text-to-Speech with Unsupervised Text Pretrainingcited by 10
  2. 2023Findings of the 2023 ML-SUPERB Challenge: Pre-Training and Evaluation over More Languages and Beyondcited by 10
  3. 2025Zero-AVSR: Zero-Shot Audio-Visual Speech Recognition with LLMs by Learning Language-Agnostic Speech Representationscited by 9
  4. 2023Unified model for code-switching speech recognition and language identification based on a concatenated tokenizercited by 8
  5. 2024Scaling A Simple Approach to Zero-Shot Speech Recognitioncited by 8
  6. 2024The Balancing Act: Unmasking and Alleviating ASR Biases in Portuguesecited by 6
  7. 2025Unveiling Biases while Embracing Sustainability: Assessing the Dual Challenges of Automatic Speech Recognition Systemscited by 4
  8. 2026PRiSM: Benchmarking Phone Realization in Speech Modelscited by 4
  9. 2024LAMA-UT: Language Agnostic Multilingual ASR through Orthography Unification and Language-Specific Transliterationcited by 3
  10. 2022Bootstrapping meaning through listening: Unsupervised learning of spoken sentence embeddingscited by 2
  11. 2024Improving Multilingual ASR in the Wild Using Simple N-best Re-rankingcited by 2
  12. 2025Improving Multilingual Speech Models on ML-SUPERB 2.0: Fine-tuning with Data Augmentation and LID-Aware CTCcited by 1
  13. 2026Hard to Be Heard: Phoneme-Level ASR Analysis of Phonologically Complex, Low-Resource Endangered Languagescited by 1
  14. 2023Unsupervised ASR via Cross-Lingual Pseudo-Labelingcited by 0