mlsanityData to 7 Oct 2026

Blank Collapse: Compressing CTC emission for the faster decoding

arXiv 2210.17017 ↗First posted 31 Oct 2022cs.CL cs.SD eess.AScited by 3

Not on the map. Why?

Abstract (arXiv)

Connectionist Temporal Classification (CTC) model is a very efficient method for modeling sequences, especially for speech data. In order to use CTC model as an Automatic Speech Recognition (ASR) task, the beam search decoding with an external language model like n-gram LM is necessary to obtain reasonable results. In this paper we analyze the blank label in CTC beam search deeply and propose a very simple method to reduce the amount of calculation resulting in faster beam search decoding speed. With this method, we can get up to 78% faster decoding speed than ordinary beam search decoding with a very small loss of accuracy in LibriSpeech datasets. We prove this method is effective not only practically by experiments but also theoretically by mathematical reasoning. We also observe that this reduction is more obvious if the accuracy of the model is higher.

Key facts. Cited by 3 papers in the map (data through 7 Oct 2026). Builds on 17 of its 31 references that are in the map. First posted on arXiv 31 Oct 2022.

Builds on

17 of 31 references.

References · 17

  1. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,160
  2. 2020wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representationscited by 3,853
  3. 2020Conformer: Convolution-augmented Transformer for Speech Recognitioncited by 1,838
  4. 2015Listen, Attend and Spellcited by 542
  5. 2017Beam Search Strategies for Neural Machine Translationcited by 232
  6. 2020Transformer Transducer: A Streamable Speech Recognition Model with Transformer Encoders and RNN-T Losscited by 232
  7. 2019End-to-end ASR: from Supervised to Semi-Supervised Learning with Modern Architecturescited by 120
  8. 2021TorchAudio: Building Blocks for Audio and Speech Processingcited by 83
  9. 2020CTC-Segmentation of Large Corpora for German End-to-end Speech Recognitioncited by 75
  10. 2015EESEN: End-to-End Speech Recognition using Deep RNN Models and WFST-based Decodingcited by 73
  11. 2014First-Pass Large Vocabulary Continuous Speech Recognition using Bi-Directional Recurrent DNNscited by 27
  12. 2022Flashlight: Enabling Innovation in Tools for Machine Learningcited by 25
5 more
  1. 2019RNN-T For Latency Controlled ASR With Improved Beam Searchcited by 19
  2. 2021Tiny Transducer: A Highly-efficient Speech Recognition Model on Edge Devicescited by 18
  3. 2021CTC Variations Through New WFST Topologiescited by 17
  4. 2022Accelerating RNN-T Training and Inference Using CTC guidancecited by 17
  5. 2018A GPU-based WFST Decoder with Exact Lattice Generationcited by 3

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 2; 2025: 0; 2026: 1.

3 papers.

  1. 2024Unveiling the Potential of LLM-Based ASR on Chinese Open-Source Datasetscited by 13
  2. 2024Guiding Frame-Level CTC Alignments Using Self-knowledge Distillationcited by 1
  3. 2026TASU2: Controllable CTC Simulation for Alignment and Low-Resource Adaptation of Speech LLMscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025FLToP CTC: Frame-Level Token Pruning via Relative Threshold for Efficient and Memory-Saving Decoding on Diverse Platforms
  2. 2026CTC-Seeded Token Edit Refinement for Non-Autoregressive Speech Recognition
  3. 2022Fast and parallel decoding for transducer
  4. 2023A CTC Alignment-based Non-autoregressive Transformer for End-to-end Automatic Speech Recognition
  5. 2022Uconv-Conformer: High Reduction of Input Sequence Length for End-to-End Speech Recognition
  6. 2024Dynamic-Width Speculative Beam Decoding for Efficient LLM Inference
  7. 2025Efficient Beam Search for Large Language Models Using Trie-Based Decoding
  8. 2025SyncSpeech: Efficient and Low-Latency Text-to-Speech based on Temporal Masked Transformer