Improving Frame-level Classifier for Word Timings with Non-peaky CTC in End-to-End Automatic Speech Recognition

Xianzhao Chen, Yist Y. Lin, Kang Wang, Yi He, Zejun Ma
6/9/2023
eess.AScs.AIcs.LG

No confident code implementation found for this paper.

Abstract

End-to-end (E2E) systems have shown comparable performance to hybrid systems for automatic speech recognition (ASR). Word timings, as a by-product of ASR, are essential in many applications, especially for subtitling and computer-aided pronunciation training. In this paper, we improve the frame-level classifier for word timings in E2E system by introducing label priors in connectionist temporal classification (CTC) loss, which is adopted from prior works, and combining low-level Mel-scale filter banks with high-level ASR encoder output as input feature. On the internal Chinese corpus, the proposed method achieves 95.68%/94.18% compared to the hybrid system 93.0%/90.22% on the word timing accuracy metrics. It also surpass a previous E2E approach with an absolute increase of 4.80%/8.02% on the metrics on 7 languages. In addition, we further improve word timing accuracy by delaying CTC peaks with frame-wise knowledge distillation, though only experimenting on LibriSpeech.

View on arXivView PDF

Code Implementations

No confident code match yet

We couldn't find an author-owned or strongly-evidenced community implementation for this paper. Any repos shown below are weak matches — verify before relying on them.

No code implementations found yet.

Know of an implementation? Let us know in the comments below!

Similar Papers

Timestamped Embedding-Matching Acoustic-to-Word CTC ASR

86% match2023

Blank-regularized CTC for Frame Skipping in Neural Transducer

81% match1 repo1.4k2023

Less Peaky and More Accurate CTC Forced Alignment by Label Priors

81% match1 repo102024

CTC-based Non-autoregressive Speech Translation

80% match1 repo122023

In-Sync: Adaptation of Speech Aware Large Language Models for ASR with Word Level Timestamp Predictions

80% match2026

EESEN: End-to-End Speech Recognition using Deep RNN Models and WFST-based Decoding

80% match1 repo2052015

Cite this paper

@article{chen2023improving,
  title  = {Improving Frame-level Classifier for Word Timings with Non-peaky CTC in End-to-End Automatic Speech Recognition},
  author = {Xianzhao Chen and Yist Y. Lin and Kang Wang and Yi He and Zejun Ma},
  year   = {2023},
  eprint = {2306.07949},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2306.07949v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2306.07949.svg)](https://codeofpaper.com/paper/2306.07949)

Discussion