Delayed Fusion: Integrating Large Language Models into First-Pass Decoding in End-to-end Speech Recognition

• URL: http://arxiv.org/abs/2501.09258v1
• Date: Thu, 16 Jan 2025 03:01:50 GMT
• Title: Delayed Fusion: Integrating Large Language Models into First-Pass Decoding in End-to-end Speech Recognition
• Authors: Takaaki Hori, Martin Kocour, Adnan Haider, Erik McDermott, Xiaodan Zhuang,
• Abstract summary: This paper presents an efficient decoding approach for end-to-end automatic speech recognition (E2E-ASR) with large language models (LLMs) We propose "delayed fusion," which applies LLM scores to ASR hypotheses with a delay during decoding. We demonstrate that delayed fusion provides improved decoding speed and accuracy compared to shallow fusion and N-best rescoring.
• Score: 17.376550014426623
• License:
• Abstract: This paper presents an efficient decoding approach for end-to-end automatic speech recognition (E2E-ASR) with large language models (LLMs). Although shallow fusion is the most common approach to incorporate language models into E2E-ASR decoding, we face two practical problems with LLMs. (1) LLM inference is computationally costly. (2) There may be a vocabulary mismatch between the ASR model and the LLM. To resolve this mismatch, we need to retrain the ASR model and/or the LLM, which is at best time-consuming and in many cases not feasible. We propose "delayed fusion," which applies LLM scores to ASR hypotheses with a delay during decoding and enables easier use of pre-trained LLMs in ASR tasks. This method can reduce not only the number of hypotheses scored by the LLM but also the number of LLM inference calls. It also allows re-tokenizion of ASR hypotheses during decoding if ASR and LLM employ different tokenizations. We demonstrate that delayed fusion provides improved decoding speed and accuracy compared to shallow fusion and N-best rescoring using the LibriHeavy ASR corpus and three public LLMs, OpenLLaMA 3B & 7B and Mistral 7B.

Related papers

• Real-time Verification and Refinement of Language Model Text Generation [60.04718679054704]
  Large language models (LLMs) have shown remarkable performance across a wide range of natural language tasks. A critical challenge remains in that they sometimes generate factually incorrect answers. We propose Streaming-VR, a novel approach designed to enhance the efficiency of verification and refinement of LLM outputs.
  arXiv  Detail & Related papers  (2025-01-14T03:59:48Z)
• SWIFT: On-the-Fly Self-Speculative Decoding for LLM Inference Acceleration [10.970637831760136]
  Speculative decoding (SD) has emerged as a widely used paradigm to accelerate the inference of large language models (LLMs) We introduce SWIFT, an on-the-fly self-speculative decoding algorithm that adaptively selects intermediate layers of LLMs to skip during inference. We show that SWIFT can achieve over a 1.3x-1.6x speedup while preserving the original distribution of the generated text.
  arXiv  Detail & Related papers  (2024-10-09T14:15:30Z)
• SALSA: Speedy ASR-LLM Synchronous Aggregation [40.91241351045586]
  We propose SALSA, which couples the decoder layers of the ASR to the LLM decoder, while synchronously advancing both decoders. We evaluate SALSA on 8 low-resource languages in the FLEURS benchmark, yielding substantial WER reductions of up to 38%.
  arXiv  Detail & Related papers  (2024-08-29T14:00:57Z)
• An Embarrassingly Simple Approach for LLM with Strong ASR Capacity [56.30595787061546]
  We focus on solving one of the most important tasks in the field of speech processing, with speech foundation encoders and large language models (LLM) Recent works have complex designs such as compressing the output temporally for the speech encoder, tackling modal alignment for the projector, and utilizing parameter-efficient fine-tuning for the LLM. We found that delicate designs are not necessary, while an embarrassingly simple composition of off-the-shelf speech encoder, LLM, and the only trainable linear projector is competent for the ASR task.
  arXiv  Detail & Related papers  (2024-02-13T23:25:04Z)
• LLMRefine: Pinpointing and Refining Large Language Models via Fine-Grained Actionable Feedback [65.84061725174269]
  Recent large language models (LLM) are leveraging human feedback to improve their generation quality. We propose LLMRefine, an inference time optimization method to refine LLM's output. We conduct experiments on three text generation tasks, including machine translation, long-form question answering (QA), and topical summarization. LLMRefine consistently outperforms all baseline approaches, achieving improvements up to 1.7 MetricX points on translation tasks, 8.1 ROUGE-L on ASQA, 2.2 ROUGE-L on topical summarization.
  arXiv  Detail & Related papers  (2023-11-15T19:52:11Z)
• Rephrase and Respond: Let Large Language Models Ask Better Questions for Themselves [57.974103113675795]
  We present a method named Rephrase and Respond' (RaR) which allows Large Language Models to rephrase and expand questions posed by humans. RaR serves as a simple yet effective prompting method for improving performance. We show that RaR is complementary to the popular Chain-of-Thought (CoT) methods, both theoretically and empirically.
  arXiv  Detail & Related papers  (2023-11-07T18:43:34Z)
• Prompting Large Language Models for Zero-Shot Domain Adaptation in Speech Recognition [33.07184218085399]
  With only a domain-specific text prompt, we propose two zero-shot ASR domain adaptation methods using LLaMA. Experiments show that, with only one domain prompt, both methods can effectively reduce word error rates (WER) on out-of-domain TedLium-2 and SPGI datasets.
  arXiv  Detail & Related papers  (2023-06-28T08:29:00Z)
• LLM-Pruner: On the Structural Pruning of Large Language Models [65.02607075556742]
  Large language models (LLMs) have shown remarkable capabilities in language understanding and generation. We tackle the compression of LLMs within the bound of two constraints: being task-agnostic and minimizing the reliance on the original training dataset. Our method, named LLM-Pruner, adopts structural pruning that selectively removes non-critical coupled structures.
  arXiv  Detail & Related papers  (2023-05-19T12:10:53Z)
• Inference with Reference: Lossless Acceleration of Large Language Models [97.04200102556551]
  LLMA is an accelerator to speed up Large Language Model (LLM) inference with references. It is motivated by the observation that there are abundant identical text spans between the decoding result by an LLM and the reference that is available in many real world scenarios.
  arXiv  Detail & Related papers  (2023-04-10T09:55:14Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.