mlsanityData to 14 Oct 2026

Improving endpoint detection in end-to-end streaming ASR for conversational speech

arXiv 2505.17070 ↗First posted 19 May 2025cs.CL cs.AI cs.SDcited by 0

Not on the map. Why?

Abstract (arXiv)

ASR endpointing (EP) plays a major role in delivering a good user experience in products supporting human or artificial agents in human-human/machine conversations. Transducer-based ASR (T-ASR) is an end-to-end (E2E) ASR modelling technique preferred for streaming. A major limitation of T-ASR is delayed emission of ASR outputs, which could lead to errors or delays in EP. Inaccurate EP will cut the user off while speaking, returning incomplete transcript while delays in EP will increase the perceived latency, degrading the user experience. We propose methods to improve EP by addressing delayed emission along with EP mistakes. To address the delayed emission problem, we introduce an end-of-word token at the end of each word, along with a delay penalty. The EP delay is addressed by obtaining a reliable frame-level speech activity detection using an auxiliary network. We apply the proposed methods on Switchboard conversational speech corpus and evaluate it against a delay penalty method.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 16 of its 25 references that are in the map. First posted on arXiv 19 May 2025.

Builds on

16 of 25 references.

References · 16

  1. 2020Conformer: Convolution-augmented Transformer for Speech Recognitioncited by 1,863
  2. 2016Joint CTC-Attention based End-to-End Speech Recognition using Multi-task Learningcited by 257
  3. 2017Online and Linear-Time Attention by Enforcing Monotonic Alignmentscited by 85
  4. 2023Zipformer: A faster and better encoder for automatic speech recognitioncited by 80
  5. 2020FastEmit: Low-latency Streaming ASR with Sequence-level Emission Regularizationcited by 69
  6. 2014End-to-end Continuous Speech Recognition using Attention-based Recurrent NN: First Resultscited by 68
  7. 2020Towards Fast and Accurate Streaming End-to-End ASRcited by 47
  8. 2020Alignment Restricted Streaming Recurrent Neural Network Transducercited by 42
  9. 2024Turn-taking and Backchannel Prediction with Acoustic and Large Language Model Fusioncited by 20
  10. 2021Dissecting User-Perceived Latency of On-Device E2E Speech Recognitioncited by 19
  11. 2021Reducing Streaming ASR Model Delay with Self Alignmentcited by 16
  12. 2022Endpoint Detection for Streaming End-to-End Multi-talker ASRcited by 10
4 more
  1. 2022Minimum Latency Training of Sequence Transducers for Streaming End-to-End Speech Recognitioncited by 8
  2. 2022Delay-penalized transducer for low-latency streaming ASRcited by 7
  3. 2022Unified End-to-End Speech Recognition and Endpointing for Fast and Efficient Speech Systemscited by 5
  4. 2021A comparison of streaming models and data augmentation methods for robust speech recognitioncited by 4

Built on by

Not cited yet.