mlsanityData to 7 Oct 2026

M-CIF: Multi-Scale Alignment For CIF-Based Non-Autoregressive ASR

arXiv 2510.22172 ↗First posted 25 Oct 2025cs.SD cs.CLcited by 0

Not on the map. Why?

Abstract (arXiv)

The Continuous Integrate-and-Fire (CIF) mechanism provides effective alignment for non-autoregressive (NAR) speech recognition. This mechanism creates a smooth and monotonic mapping from acoustic features to target tokens, achieving performance on Mandarin competitive with other NAR approaches. However, without finer-grained guidance, its stability degrades in some languages such as English and French. In this paper, we propose Multi-scale CIF (M-CIF), which performs multi-level alignment by integrating character and phoneme level supervision progressively distilled into subword representations, thereby enhancing robust acoustic-text alignment. Experiments show that M-CIF reduces WER compared to the Paraformer baseline, especially on CommonVoice by 4.21% in German and 3.05% in French. To further investigate these gains, we define phonetic confusion errors (PE) and space-related segmentation errors (SE) as evaluation metrics. Analysis of these metrics across different M-CIF settings reveals that the phoneme and character layers are essential for enhancing progressive CIF alignment.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 21 of its 32 references that are in the map. First posted on arXiv 25 Oct 2025.

Builds on

21 of 32 references.

References · 21

  1. 2017Attention Is All You Needcited by 59,932
  2. 2018SentencePiece: A simple and language independent subword tokenizer and detokenizer for Neural Text Processingcited by 1,973
  3. 2020Conformer: Convolution-augmented Transformer for Speech Recognitioncited by 1,838
  4. 2019Common Voice: A Massively-Multilingual Speech Corpuscited by 1,358
  5. 2019SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognitioncited by 1,355
  6. 2017AISHELL-1: An Open-Source Mandarin Speech Corpus and A Speech Recognition Baselinecited by 364
  7. 2018Deterministic Non-Autoregressive Neural Sequence Modeling by Iterative Refinementcited by 194
  8. 2022Paraformer: Fast and Accurate Parallel Transformer for Non-autoregressive End-to-End Speech Recognitioncited by 139
  9. 2021Recent Advances in End-to-End Automatic Speech Recognitioncited by 138
  10. 2023FunASR: A Fundamental End-to-End Speech Recognition Toolkitcited by 92
  11. 2019CIF: Continuous Integrate-and-Fire for End-to-End Speech Recognitioncited by 76
  12. 2020Mask CTC: Non-Autoregressive End-to-End ASR with CTC and Mask Predictcited by 75
9 more
  1. 2020Imputer: Sequence Modelling via Imputation and Dynamic Programmingcited by 65
  2. 2020Align-Refine: Non-Autoregressive Speech Recognition via Iterative Realignmentcited by 37
  3. 2020CASS-NAT: CTC Alignment-based Single Step Non-autoregressive Transformer for Speech Recognitioncited by 24
  4. 2021Boundary and Context Aware Training for CIF-based Non-Autoregressive End-to-end ASRcited by 12
  5. 2023Knowledge Transfer from Pre-trained Language Models to Cif-based Speech Recognizers via Hierarchical Distillationcited by 11
  6. 2023CIF-T: A Novel CIF-based Transducer Architecture for Automatic Speech Recognitioncited by 6
  7. 2025Soundwave: Less is More for Speech-Text Alignment in LLMscited by 6
  8. 2023CIF-PT: Bridging Speech and Text Representations for Spoken Language Understanding via Continuous Integrate-and-Fire Pre-Trainingcited by 3
  9. 2024An efficient text augmentation approach for contextualized Mandarin speech recognitioncited by 2

Built on by

Not cited yet.