mlsanityData to 19 Oct 2026

Speech recognition models and training

1,811 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: asr, speech, recognition, speech recognition, language

Key facts. Speech recognition models and training: 1,811 papers on arXiv since 2021 (data through 19 Oct 2026). 448 of its 1,811 mapped papers (25%) cite “Robust Speech Recognition via Large-Scale Weak Supervision” (2022, arXiv 2212.04356). 20 new papers 2 Oct–19 Oct.

What’s new

2 Oct – 19 Oct 2026

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

  1. This week · 3
  2. Preference-Aware Pareto Steering for Multi-Task Speech LLMs also fits: Speech translation
  3. Think Before You Transcribe: Context-Steered Reasoning for Hotword-Aware Speech Recognition
  4. PronCFT: Pronunciation Contextual Fine-Tuning for Contextual Biasing of Out-of-Vocabulary Words in LLM-Based ASR also fits: ASR evaluation, bias and atypical speech
  5. Earlier · 17
  6. Ask, Don't Guess: Targeted Clarification for Speech Recognition Errors
  7. Flash RNN-T: Memory-Efficient Joint Network and Transducer Loss Computation
  8. Zero-Shot Spoken Language Understanding with SLAM-ASR: What Do Monolingual ASR-Trained Linear Projectors Enable? also fits: Multilingual and low-resource ASR
  9. Towards Efficient Simultaneous Inverse Text Normalization with Pretrained Text-to-Text Language Model and Read-Tag-Write Policy also fits: Speech translation
  10. SSEmb: Training a Generalizable Speech Retrieval Model Using Only Automatic Speech Recognition Data
  11. Towards Privacy-Preserving ASR: Speaker-Level Machine Unlearning also fits: Anti-spoofing and adversarial audio attacks
  12. Select or Rewrite: Oracle-Guided Fine-Tuning for Generative ASR Post-Editing from N-Best Hypotheses also fits: Audio language models and spoken dialogue
All 20
  1. ASSET: Adaptive Speech Segmentation via Reinforcement Learning
  2. Lossless Inference Acceleration for ASR via EAGLE-Style Speculative Decoding
  3. O'ido: Open-Vocabulary Speech Recognition on a 5 Microcontroller without a Neural Accelerator also fits: Multilingual and low-resource ASR
  4. Does Financial Speech Need Continual Learning Regularization? A Study of EWC and mini-KFAC on Earnings Call ASR also fits: ASR evaluation, bias and atypical speech
  5. From 2.1M Hours of Raw Broadcast Audio to Finnish Speech SSL: Confidence Filtering and Architectural Analysis at Scale also fits: Multilingual and low-resource ASR
  6. When Can You Prune Your Network? A Study of Intermediate Neurons in Multilingual Speech Parsing also fits: Multilingual and low-resource ASR
  7. DirectSpeech2LLM: A Simple End-to-End Framework to Mitigate Prompt Overfitting in Speech-LLMs
  8. HINTT Submission to the 2nd MLC-SLM Challenge: Comparing Cascaded and Unified Approaches to Diarization and ASR also fits: Multilingual and low-resource ASR
  9. Lyric: Wave-Domain Computing for Efficient Spoken-Digit Recognition also fits: Speaker recognition and diarization
  10. Factorized Delayed Streams Modeling for LLM-based Streaming ASR

Foundations

Most cited by papers here

  1. Robust Speech Recognition via Large-Scale Weak Supervision2022cited by 448 here
  2. WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing2021cited by 242 here
  3. Recent Advances in End-to-End Automatic Speech Recognition2021cited by 95 here
  4. Contextualized Streaming End-to-End Speech Recognition with Trie-Based Deep Biasing and Shallow Fusion2021cited by 94 here
  5. WeNet: Production oriented Streaming and Non-streaming End-to-End Speech Recognition Toolkit2021cited by 86 here

Lineage

The papers this direction builds on, oldest first.

Each paper on the line cites the one before.

  1. 2021XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scalecited by 81 here
  2. 2022FLEURS: Few-shot Learning Evaluation of Universal Representations of Speechcited by 79 here
  3. 2022Robust Speech Recognition via Large-Scale Weak Supervisioncited by 448 here
  4. 2023Google USM: Scaling Automatic Speech Recognition Beyond 100 Languagescited by 72 here
  5. 2023AudioPaLM: A Large Language Model That Can Speak and Listencited by 64 here
  6. 2023SALMONN: Towards Generic Hearing Abilities for Large Language Modelscited by 77 here
18 more it builds on, not in this chain
  1. 2021Generative Spoken Language Modeling from Raw Audiocited by 57 here
  2. 2021WeNet: Production oriented Streaming and Non-streaming End-to-End Speech Recognition Toolkitcited by 86 here
  3. 2021Intermediate Loss Regularization for CTC-based Speech Recognitioncited by 67 here
  4. 2021Contextualized Streaming End-to-End Speech Recognition with Trie-Based Deep Biasing and Shallow Fusioncited by 94 here
  5. 2021GigaSpeech: An Evolving, Multi-domain ASR Corpus with 10,000 Hours of Transcribed Audiocited by 120 here
  6. 2021Layer-wise Analysis of a Self-supervised Speech Representation Modelcited by 67 here
  7. 2021WenetSpeech: A 10000+ Hours Multi-domain Mandarin Corpus for Speech Recognitioncited by 58 here
  8. 2021WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processingcited by 242 here
  9. 2021Recent Advances in End-to-End Automatic Speech Recognitioncited by 95 here
  10. 2021Context-Aware Transformer Transducer for Speech Recognitioncited by 55 here
  11. 2022Self-supervised Learning with Random-projection Quantizer for Speech Recognitioncited by 63 here
  12. 2022Self-Supervised Speech Representation Learning: A Reviewcited by 74 here
  13. 2022Contextual Adapters for Personalized Speech Recognition in Neural Transducerscited by 54 here
  14. 2022E-Branchformer: Branchformer with Enhanced merging for speech recognitioncited by 55 here
  15. 2023End-to-End Speech Recognition: A Surveycited by 71 here
  16. 2023Scaling Speech Technology to 1,000+ Languagescited by 63 here
  17. 2023Prompting Large Language Models with Speech Recognition Abilitiescited by 65 here
  18. 2023Qwen-Audio: Advancing Universal Audio Understanding via Unified Large-Scale Audio-Language Modelscited by 62 here

The 24 most cited of 2,312.

Browse

30 of 1,811, ranked 6–35

  1. 6Self-Supervised Speech Representation Learning: A Review202274
  2. 7End-to-End Speech Recognition: A Survey202371
  3. 8Intermediate Loss Regularization for CTC-based Speech Recognition202167
  4. 9Layer-wise Analysis of a Self-supervised Speech Representation Model202167
  5. 10Prompting Large Language Models with Speech Recognition Abilities202365
  6. 11Self-supervised Learning with Random-projection Quantizer for Speech Recognition202263
  7. 12E-Branchformer: Branchformer with Enhanced merging for speech recognition202255
  8. 13Context-Aware Transformer Transducer for Speech Recognition202155
  9. 14Generative Speech Recognition Error Correction with Large Language Models and Task-Activating Prompting202354
  10. 15Contextual Adapters for Personalized Speech Recognition in Neural Transducers202254
All 30
  1. 16Unsupervised Speech Recognition202152
  2. 17SpeechT5: Unified-Modal Encoder-Decoder Pre-Training for Spoken Language Processing202152
  3. 18Fast Conformer with Linearly Scalable Attention for Efficient Speech Recognition202350
  4. 19Relaxing the Conditional Independence Assumption of CTC-based ASR by Conditioning on Intermediate Predictions202148
  5. 20An Embarrassingly Simple Approach for LLM with Strong ASR Capacity202447
  6. 21Can Generative Large Language Models Perform ASR Error Correction?202346
  7. 22MAESTRO: Matched Speech Text Representations through Modality Matching202246
  8. 23SALM: Speech-augmented Language Model with In-context Learning for Speech Recognition and Translation202343
  9. 24Branchformer: Parallel MLP-Attention Architectures to Capture Local and Global Context for Speech Recognition and Understanding202243
  10. 25SLUE: New Benchmark Tasks for Spoken Language Understanding Evaluation on Natural Speech202143
  11. 26ESPnet-SLU: Advancing Spoken Language Understanding through ESPnet202142
  12. 27Zipformer: A faster and better encoder for automatic speech recognition202341
  13. 28WeNet 2.0: More Productive End-to-End Speech Recognition Toolkit202241
  14. 29Internal Language Model Training for Domain-Adaptive End-to-End Speech Recognition202141
  15. 30Paraformer: Fast and Accurate Parallel Transformer for Non-autoregressive End-to-End Speech Recognition202240
  16. 31N-best T5: Robust ASR Error Correction using Multiple Input Hypotheses and Constrained Decoding Space202339
  17. 32SLAM: A Unified Encoder for Speech and Language Modeling via Speech-Text Joint Pre-Training202137
  18. 33Prompting Large Language Models for Zero-Shot Domain Adaptation in Speech Recognition202336
  19. 34HyPoradise: An Open Baseline for Generative Speech Recognition with Large Language Models202336
  20. 35Improving End-to-End Contextual Speech Recognition with Fine-Grained Contextual Knowledge Selection202236