Distilling an End-to-End Voice Assistant Without Instruction Training Data

• URL: http://arxiv.org/abs/2410.02678v1
• Date: Thu, 3 Oct 2024 17:04:48 GMT
• Title: Distilling an End-to-End Voice Assistant Without Instruction Training Data
• Authors: William Held, Ella Li, Michael Ryan, Weiyan Shi, Yanzhe Zhang, Diyi Yang,
• Abstract summary: Distilled Voice Assistant (DiVA) generalizes to Question Answering, Classification, and Translation. We show that DiVA better meets user preferences, achieving a 72% win rate compared with state-of-the-art models like Qwen 2 Audio.
• Score: 53.524071162124464
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Voice assistants, such as Siri and Google Assistant, typically model audio and text separately, resulting in lost speech information and increased complexity. Recent efforts to address this with end-to-end Speech Large Language Models (LLMs) trained with supervised finetuning (SFT) have led to models ``forgetting" capabilities from text-only LLMs. Our work proposes an alternative paradigm for training Speech LLMs without instruction data, using the response of a text-only LLM to transcripts as self-supervision. Importantly, this process can be performed without annotated responses. We show that our Distilled Voice Assistant (DiVA) generalizes to Spoken Question Answering, Classification, and Translation. Furthermore, we show that DiVA better meets user preferences, achieving a 72\% win rate compared with state-of-the-art models like Qwen 2 Audio, despite using $>$100x less training compute.

Related papers

• VoiceLDM: Text-to-Speech with Environmental Context [22.29992463094861]
  VoiceLDM is a model designed to produce audio that accurately follows two distinct natural language text prompts. By utilizing pretrained contrastive language-audio pretraining (CLAP) and Whisper, VoiceLDM is trained on large amounts of real-world audio without manual annotations or transcriptions. We show that VoiceLDM is capable of generating plausible audio that aligns well with both input conditions, even surpassing the speech intelligibility of the ground truth audio on the AudioCaps test set.
  arXiv  Detail & Related papers  (2023-09-24T15:20:59Z)
• Weakly-supervised Automated Audio Captioning via text only training [1.504795651143257]
  We propose a weakly-supervised approach to train an AAC model assuming only text data and a pre-trained CLAP model. We evaluate our proposed method on Clotho and AudioCaps datasets demonstrating its ability to achieve a relative performance of up to $83%$ compared to fully supervised approaches.
  arXiv  Detail & Related papers  (2023-09-21T16:40:46Z)
• End-to-End Speech Recognition Contextualization with Large Language Models [25.198480789044346]
  We introduce a novel method for contextualizing speech recognition models incorporating Large Language Models (LLMs) We provide audio features, along with optional text tokens for context, to train the system to complete transcriptions in a decoder-only fashion. Our empirical results demonstrate a significant improvement in performance, with a 6% WER reduction when additional textual context is provided.
  arXiv  Detail & Related papers  (2023-09-19T20:28:57Z)
• AudioPaLM: A Large Language Model That Can Speak and Listen [79.44757696533709]
  We introduce AudioPaLM, a large language model for speech understanding and generation. AudioPaLM fuses text-based and speech-based language models. It can process and generate text and speech with applications including speech recognition and speech-to-speech translation.
  arXiv  Detail & Related papers  (2023-06-22T14:37:54Z)
• Spoken Question Answering and Speech Continuation Using Spectrogram-Powered LLM [19.36630667212398]
  We present Spectron, a novel approach to adapting pre-trained large language models (LLMs) to perform spoken question answering (QA) and speech continuation. Key to our approach is a training objective that jointly supervises speech recognition, text continuation, and speech synthesis. Our method surpasses existing spoken language models in speaker preservation and semantic coherence.
  arXiv  Detail & Related papers  (2023-05-24T15:39:43Z)
• SpeechUT: Bridging Speech and Text with Hidden-Unit for Encoder-Decoder Based Speech-Text Pre-training [106.34112664893622]
  We propose a unified-modal speech-unit-text pre-training model, SpeechUT, to connect the representations of a speech encoder and a text decoder with a shared unit encoder. Our proposed SpeechUT is fine-tuned and evaluated on automatic speech recognition (ASR) and speech translation (ST) tasks.
  arXiv  Detail & Related papers  (2022-10-07T17:57:45Z)
• WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing [102.45426364965887]
  We propose a new pre-trained model, WavLM, to solve full-stack downstream speech tasks. WavLM is built based on the HuBERT framework, with an emphasis on both spoken content modeling and speaker identity preservation. We scale up the training dataset from 60k hours to 94k hours of public audio data, and optimize its training procedure for better representation extraction.
  arXiv  Detail & Related papers  (2021-10-26T17:55:19Z)
• Zero-Shot Text-to-Speech for Text-Based Insertion in Audio Narration [62.75234183218897]
  We propose a one-stage context-aware framework to generate natural and coherent target speech without any training data of the speaker. We generate the mel-spectrogram of the edited speech with a transformer-based decoder. It outperforms a recent zero-shot TTS engine by a large margin.
  arXiv  Detail & Related papers  (2021-09-12T04:17:53Z)
• Exploring Transfer Learning For End-to-End Spoken Language Understanding [8.317084844841323]
  An end-to-end (E2E) system that goes directly from speech to a hypothesis is a more attractive option. We propose an E2E system that is designed to jointly train on multiple speech-to-text tasks. We show that it beats the performance of E2E models trained on individual tasks.
  arXiv  Detail & Related papers  (2020-12-15T19:02:15Z)
• Audio ALBERT: A Lite BERT for Self-supervised Learning of Audio Representation [51.37980448183019]
  We propose Audio ALBERT, a lite version of the self-supervised speech representation model. We show that Audio ALBERT is capable of achieving competitive performance with those huge models in the downstream tasks. In probing experiments, we find that the latent representations encode richer information of both phoneme and speaker than that of the last layer.
  arXiv  Detail & Related papers  (2020-05-18T10:42:44Z)

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.