mlsanityData to 19 Oct 2026

Multilingual and low-resource ASR

1,075 papers3 this weekMedium confidence in this grouping: mostly one topic, with blurry edgesMedium confidence in this grouping: mostly one topic, with blurry edges · About

Keywords: speech, asr, languages, recognition, language

Key facts. Multilingual and low-resource ASR: 1,075 papers on arXiv since 2021 (data through 19 Oct 2026). 171 of its 1,075 mapped papers (16%) cite “XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scale” (2021, arXiv 2111.09296). 18 new papers 2 Oct–19 Oct.

What’s new

2 Oct – 19 Oct 2026

18 new papers, 2 Oct – 19 Oct; 3 this week.

  1. This week · 3
  2. Evaluating Tibetan Speech Recognition Across Dialects and Adaptation Regimes: A Multi-Granularity Analysis
  3. Monsoon: Scalable Spontaneous Speech Collection for Low-Resource ASR in 50 Languages also fits: Speech recognition models and training
  4. TeleASR-Amharic: A Benchmark for Robust Amharic ASR in Realistic Telecommunication Conditions also fits: Speech recognition models and training
  5. Earlier · 15
  6. Anuvani-LID: Robust and Calibrated Spoken Language Identification for Indian Languages
  7. SEA-HELM Audio: Evaluating Multilingual ASR-LLM Pipeline Performance Across SEA-HELM Subsets
  8. ELARSpeech: A 30K-Hour Dataset Across 675 Underrepresented Languages for Long-Tail Multilingual Speech Technologies also fits: Speech translation
  9. Mixture of Phonetic Experts Based Low-Rank Adaptation of Conformer Models for Accented English Speech Recognition
  10. CAFE: Confining Feature Distillation to Language-Routed Adapters for Compact Multilingual Automatic Speech Recognition also fits: Speech recognition models and training
  11. Code-Switching Detection via Language-Selective Dimensions in Speech Foundation Models
  12. Nufo: A Benchmark for Ewe Automatic Speech Recognition
All 18
  1. LLM-HB: Language-Aware LLM-Guided Hotword Biasing for Code-Switching ASR also fits: Speech recognition models and training
  2. Phonologically Informed Tokenization for German Speech Recognition: A Cross-Domain Study
  3. Itgan at NADI 2026 shared task: Parameter-Efficient Whisper Adaptation for Robust, Mixed-Dialect and Code-Switched Arabic ASR
  4. Mitigating Accent-Language Confusion in Self-Supervised Speech Representations for Language Identification
  5. Phoneme-Guided Initialization for LLM-based Speech Recognition also fits: Speech recognition models and training
  6. Automatic Speech Recognition for Low-Resource Sinhala: A Critical Review of Methods, Challenges, and Future Directions
  7. Offline AI Modules: Voice-First Offline Architecture, Hardware Reference Stack, Quantization and Benchmarking also fits: Audio language models and spoken dialogue
  8. AraYoungVoices: A Diverse L1/L2 Corpus of Arabic Child and Adolescent Speech also fits: ASR evaluation, bias and atypical speech

Foundations

Most cited by papers here

  1. XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scale2021cited by 171 here
  2. Scaling Speech Technology to 1,000+ Languages2023cited by 170 here
  3. FLEURS: Few-shot Learning Evaluation of Universal Representations of Speech2022cited by 164 here
  4. Google USM: Scaling Automatic Speech Recognition Beyond 100 Languages2023cited by 88 here
  5. GigaSpeech: An Evolving, Multi-domain ASR Corpus with 10,000 Hours of Transcribed Audio2021cited by 48 here

Lineage

The papers this direction builds on, oldest first.

Each paper on the line cites the one before.

  1. 2021Scaling End-to-End Models for Large-Scale Multilingual ASRcited by 41 here
  2. 2021XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scalecited by 171 here
  3. 2022FLEURS: Few-shot Learning Evaluation of Universal Representations of Speechcited by 164 here
  4. 2022Robust Speech Recognition via Large-Scale Weak Supervisioncited by 456 here
  5. 2023Google USM: Scaling Automatic Speech Recognition Beyond 100 Languagescited by 88 here
  6. 2023Scaling Speech Technology to 1,000+ Languagescited by 170 here
18 more it builds on, not in this chain
  1. 2021WeNet: Production oriented Streaming and Non-streaming End-to-End Speech Recognition Toolkitcited by 27 here
  2. 2021Multilingual and code-switching ASR challenges for low resource Indian languagescited by 32 here
  3. 2021Unsupervised Speech Recognitioncited by 26 here
  4. 2021Towards One Model to Rule All: Multilingual Strategy for Dialectal Code-Switching Arabic ASRcited by 28 here
  5. 2021Conditional Variational Autoencoder with Adversarial Learning for End-to-End Text-to-Speechcited by 56 here
  6. 2021GigaSpeech: An Evolving, Multi-domain ASR Corpus with 10,000 Hours of Transcribed Audiocited by 48 here
  7. 2021QASR: QCRI Aljazeera Speech Resource -- A Large Scale Annotated Arabic Speech Corpuscited by 31 here
  8. 2021A Configurable Multilingual Model is All You Need to Recognize All Languagescited by 26 here
  9. 2021WenetSpeech: A 10000+ Hours Multi-domain Mandarin Corpus for Speech Recognitioncited by 44 here
  10. 2021WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processingcited by 94 here
  11. 2021Towards Building ASR Systems for the Next Billion Userscited by 30 here
  12. 2022Reducing Language confusion for Code-switching Speech Recognition with Token-level Language Diarizationcited by 26 here
  13. 2023Neural Codec Language Models are Zero-Shot Text to Speech Synthesizerscited by 35 here
  14. 2023Fast Conformer with Linearly Scalable Attention for Efficient Speech Recognitioncited by 32 here
  15. 2023SeamlessM4T: Massively Multilingual & Multimodal Machine Translationcited by 28 here
  16. 2023Seamless: Multilingual Expressive and Streaming Speech Translationcited by 40 here
  17. 2024XTTS: a Massively Multilingual Zero-Shot Text-to-Speech Modelcited by 29 here
  18. 2024Qwen2-Audio Technical Reportcited by 29 here

The 24 most cited of 1,797.

Browse

30 of 1,075, ranked 6–35

  1. 6WenetSpeech: A 10000+ Hours Multi-domain Mandarin Corpus for Speech Recognition202144
  2. 7Scaling End-to-End Models for Large-Scale Multilingual ASR202141
  3. 8Multilingual and code-switching ASR challenges for low resource Indian languages202132
  4. 9QASR: QCRI Aljazeera Speech Resource -- A Large Scale Annotated Arabic Speech Corpus202131
  5. 10Towards Building ASR Systems for the Next Billion Users202130
  6. 11Towards One Model to Rule All: Multilingual Strategy for Dialectal Code-Switching Arabic ASR202128
  7. 12Reducing Language confusion for Code-switching Speech Recognition with Token-level Language Diarization202226
  8. 13A Configurable Multilingual Model is All You Need to Recognize All Languages202126
  9. 14WhisperX: Time-Accurate Speech Transcription of Long-Form Audio202325
  10. 15IndicSUPERB: A Speech Processing Universal Performance Benchmark for Indian languages202224
All 30
  1. 16LAE: Language-Aware Encoder for Monolingual and Multilingual ASR202223
  2. 17LoRA-Whisper: Parameter-Efficient and Extensible Multilingual ASR202422
  3. 18Language-specific Characteristic Assistance for Code-switching Speech Recognition202222
  4. 19The Multilingual TEDx Corpus for Speech Recognition and Translation202120
  5. 20Omnilingual ASR: Open-Source Multilingual Speech Recognition for 1600+ Languages202519
  6. 21Improving Massively Multilingual ASR With Auxiliary CTC Objectives202319
  7. 22Language-Routing Mixture of Experts for Multilingual and Code-Switching Speech Recognition202319
  8. 23GigaSpeech 2: An Evolving, Large-Scale and Multi-domain ASR Corpus for Low-Resource Languages with Automated Crawling, Transcription and Refinement202418
  9. 24UniSpeech: Unified Speech Representation Learning with Labeled and Unlabeled Data202118
  10. 25Exploiting Adapters for Cross-lingual Low-resource Speech Recognition202118
  11. 26Joint Unsupervised and Supervised Training for Multilingual ASR202118
  12. 27The People's Speech: A Large-Scale Diverse English Speech Recognition Dataset for Commercial Usage202118
  13. 28Joint Modeling of Code-Switched and Monolingual ASR via Conditional Factorization202118
  14. 29Distil-Whisper: Robust Knowledge Distillation via Large-Scale Pseudo Labelling202317
  15. 30Towards Zero-Shot Code-Switched Speech Recognition202217
  16. 31Making More of Little Data: Improving Low-Resource Automatic Speech Recognition Using Data Augmentation202316
  17. 32Vistaar: Diverse Benchmarks and Training Sets for Indian Language ASR202316
  18. 33N-Shot Benchmarking of Whisper on Diverse Arabic Speech Recognition202316
  19. 34ASR2K: Speech Recognition for Around 2000 Languages without Audio202216
  20. 35The Accented English Speech Recognition Challenge 2020: Open Datasets, Tracks, Baselines, Results and Methods202116