mlsanityData to 14 Oct 2026

Updated Corpora and Benchmarks for Long-Form Speech Recognition

arXiv 2309.15013 ↗First posted 26 Sep 2023cs.CL cs.SD eess.AScited by 10

Not on the map. Why?

Abstract (arXiv)

The vast majority of ASR research uses corpora in which both the training and test data have been pre-segmented into utterances. In most real-word ASR use-cases, however, test audio is not segmented, leading to a mismatch between inference-time conditions and models trained on segmented utterances. In this paper, we re-release three standard ASR corpora - TED-LIUM 3, Gigapeech, and VoxPopuli-en - with updated transcription and alignments to enable their use for long-form ASR research. We use these reconstituted corpora to study the train-test mismatch problem for transducers and attention-based encoder-decoders (AEDs), confirming that AEDs are more susceptible to this issue. Finally, we benchmark a simple long-form training for these models, showing its efficacy for model robustness under this domain shift.

Key facts. Cited by 10 papers in the map (data through 14 Oct 2026). Builds on 29 of its 41 references that are in the map. First posted on arXiv 26 Sep 2023.

Builds on

29 of 41 references.

References · 29

  1. 2022Robust Speech Recognition via Large-Scale Weak Supervisioncited by 4,057
  2. 2020Conformer: Convolution-augmented Transformer for Speech Recognitioncited by 1,863
  3. 2019SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognitioncited by 1,380
  4. 2021VoxPopuli: A Large-Scale Multilingual Speech Corpus for Representation Learning, Semi-Supervised Learning and Interpretationcited by 463
  5. 2023Scaling Speech Technology to 1,000+ Languagescited by 411
  6. 2015Attention-Based Models for Speech Recognitioncited by 401
  7. 2021GigaSpeech: An Evolving, Multi-domain ASR Corpus with 10,000 Hours of Transcribed Audiocited by 399
  8. 2016Joint CTC-Attention based End-to-End Speech Recognition using Multi-task Learningcited by 257
  9. 2018TED-LIUM 3: twice as much data and corpus repartition for experiments on speaker adaptationcited by 226
  10. 2019Recognizing long-form speech using streaming end-to-end modelscited by 77
  11. 2021Earnings-21: A Practical Benchmark for ASR in the Wildcited by 62
  12. 2022Earnings-22: A Practical Benchmark for Accents in the Wildcited by 50
17 more
  1. 2022Pruned RNN-T for fast, memory-efficient ASR trainingcited by 46
  2. 2019A comparison of end-to-end models for long-form speech recognitioncited by 42
  3. 2021Lhotse: a speech data representation library for the modern deep learning ecosystemcited by 33
  4. 2018Dialog-context aware end-to-end speech recognitioncited by 27
  5. 2020RNN-T Models Fail to Generalize to Out-of-Domain Audio: Causes and Solutionscited by 26
  6. 2020End-to-End Automatic Speech Recognition Integrated With CTC-Based Voice Activity Detectioncited by 20
  7. 2022E2E Segmenter: Joint Segmenting and Decoding for Long-Form ASRcited by 17
  8. 2023Large-scale Language Model Rescoring on Long-form Datacited by 14
  9. 2021Input Length Matters: Improving RNN-T and MWER Training for Long-form Telephony Speech Recognitioncited by 10
  10. 2023Bypass Temporal Classification: Weakly Supervised Automatic Speech Recognition with Imperfect Transcriptscited by 10
  11. 2020Improving RNN-T ASR Accuracy Using Context Audiocited by 8
  12. 2022Leveraging Acoustic Contextual Representation by Audio-textual Cross-modal Learning for Conversational ASRcited by 7
  13. 2022LongFNT: Long-form Speech Recognition with Factorized Neural Transducercited by 6
  14. 2023Towards Effective and Compact Contextual Representation for Conformer Transducer Speech Recognition Systemscited by 6
  15. 2020LSTM-LM with Long-Term History for First-Pass Decoding in Conversational Speech Recognitioncited by 4
  16. 2022Random Utterance Concatenation Based Data Augmentation for Improving Short-video Speech Recognitioncited by 3
  17. 2020Context Dependent RNNLM for Automatic Transcription of Conversationscited by 2

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 4; 2025: 2; 2026: 4.

10 papers.

  1. 2026Nemotron 3 Nano Omni: Efficient and Open Multimodal Intelligencecited by 30
  2. 2024On the Effects of Heterogeneous Data Sources on Speech-to-Text Foundation Modelscited by 10
  3. 2024Longer is (Not Necessarily) Stronger: Punctuated Long-Sequence Training for Enhanced Speech Recognition and Translationcited by 4
  4. 2026Whisper-CD: Accurate Long-Form Speech Recognition using Multi-Negative Contrastive Decodingcited by 3
  5. 2024Exploring SSL Discrete Speech Features for Zipformer-based Contextual ASRcited by 2
  6. 2024Efficiently Identifying Low-Quality Language Subsets in Multilingual Datasets: A Case Study on a Large-Scale Multilingual Audio Datasetcited by 1
  7. 2025Accurate, fast, cheap: Choose three. Replacing Multi-Head-Attention with Bidirectional Recurrent Attention for Long-Form ASRcited by 0
  8. 2025Segmental Attention Decoding With Long Form Acoustic Encodingscited by 0
  9. 2026MURMUR: An Efficient Inference System for Long-Form ASRcited by 0
  10. 2026FBK's Long-form SpeechLLMs for IWSLT 2026 Instruction Followingcited by 0