Papers
Topics
Authors
Recent
Detailed Answer
Quick Answer
Concise responses based on abstracts only
Detailed Answer
Well-researched responses based on abstracts and relevant paper content.
Custom Instructions Pro
Preferences or requirements that you'd like Emergent Mind to consider when generating responses
Gemini 2.5 Flash
Gemini 2.5 Flash 64 tok/s
Gemini 2.5 Pro 50 tok/s Pro
GPT-5 Medium 30 tok/s Pro
GPT-5 High 35 tok/s Pro
GPT-4o 77 tok/s Pro
Kimi K2 174 tok/s Pro
GPT OSS 120B 457 tok/s Pro
Claude Sonnet 4 37 tok/s Pro
2000 character limit reached

Promptformer: Prompted Conformer Transducer for ASR (2401.07360v1)

Published 14 Jan 2024 in cs.CL, cs.SD, and eess.AS

Abstract: Context cues carry information which can improve multi-turn interactions in automatic speech recognition (ASR) systems. In this paper, we introduce a novel mechanism inspired by hyper-prompting to fuse textual context with acoustic representations in the attention mechanism. Results on a test set with multi-turn interactions show that our method achieves 5.9% relative word error rate reduction (rWERR) over a strong baseline. We show that our method does not degrade in the absence of context and leads to improvements even if the model is trained without context. We further show that leveraging a pre-trained sentence-piece model for context embedding generation can outperform an external BERT model.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (24)
  1. Contextual-utterance training for automatic speech recognition. In iberSPEECH, 2022.
  2. Joint modelling of spoken language understanding tasks with integrated dialog history. In ICASSP, 2023.
  3. Dialog act guided contextual adapter for personalized speech recognition. In ICASSP, 2023.
  4. Context-aware transformer transducer for speech recognition. In ASRU, 2021.
  5. Context-aware end-to-end asr using self-attentive embedding and tensor fusion. In ICASSP, 2023.
  6. BERT: pre-training of deep bidirectional transformers for language understanding. CoRR, 2018.
  7. A. Graves. Sequence transduction with recurrent neural networks. In ICML, Edinburgh, Scotland, 2012.
  8. Conformer: Convolution-augmented transformer for speech recognition. ArXiv, 2005.08100, 2020.
  9. Hyperprompt: Prompt-based task-conditioning of transformers. In PMLR, 2022.
  10. Advanced long-context end-to-end speech recognition using context-expanded transformers. In Interspeech, 2021.
  11. Contextual rnn-t for open domain asr. arXiv preprint arXiv:2006.03411, 2020.
  12. D. P. Kingma and J. Ba. Adam: A method for stochastic optimization. In ICLR, 2015.
  13. On the sentence embeddings from pre-trained language models. In EMNLP. Association for Computational Linguistics, 2020.
  14. P-tuning: Prompt tuning can be comparable to fine-tuning across scales and tasks. In ACL, Dublin, Ireland, 2022.
  15. Hierarchical transformer-based large-context end-to-end asr with large-context knowledge distillation. In ICASSP, 2021.
  16. Procter: Pronunciation-aware contextual adapter for personalized speech recognition in neural transducers. In ICASSP, 2023.
  17. SpecAugment: A simple data augmentation method for automatic speech recognition. In Interspeech. ISCA, 2019.
  18. Deep context: end-to-end contextual speech recognition. In IEEE SLT, 2018.
  19. Contextual adapters for personalized speech recognition in neural transducers. In ICASSP, 2022.
  20. Towards end-to-end integration of dialog history for improved spoken language understanding. In ICASSP.
  21. Adaptive global-local context fusion for multi-turn spoken language understanding. AAAI, 2022.
  22. Improving transformer-based conversational asr by inter-sentential attention mechanism. arXiv preprint arXiv:2207.00883, 2022.
  23. Attentive contextual carryover for multi-turn end-to-end spoken language understanding. In ASRU, 2021.
  24. Pushing the limits of semi-supervised learning for automatic speech recognition. CoRR, 2020.
Citations (3)
List To Do Tasks Checklist Streamline Icon: https://streamlinehq.com

Collections

Sign up for free to add this paper to one or more collections.

Summary

We haven't generated a summary for this paper yet.

Dice Question Streamline Icon: https://streamlinehq.com

Follow-Up Questions

We haven't generated follow-up questions for this paper yet.