Align-SLM: Textless Spoken Language Models with Reinforcement Learning from AI Feedback

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Lin, Guan-Ting, Shivakumar, Prashanth Gurunath, Gourav, Aditya, Gu, Yile, Gandhe, Ankur, Lee, Hung-yi, Bulyko, Ivan
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913860269637632
author Lin, Guan-Ting
Shivakumar, Prashanth Gurunath
Gourav, Aditya
Gu, Yile
Gandhe, Ankur
Lee, Hung-yi
Bulyko, Ivan
author_facet Lin, Guan-Ting
Shivakumar, Prashanth Gurunath
Gourav, Aditya
Gu, Yile
Gandhe, Ankur
Lee, Hung-yi
Bulyko, Ivan
contents While textless Spoken Language Models (SLMs) have shown potential in end-to-end speech-to-speech modeling, they still lag behind text-based Large Language Models (LLMs) in terms of semantic coherence and relevance. This work introduces the Align-SLM framework, which leverages preference optimization inspired by Reinforcement Learning with AI Feedback (RLAIF) to enhance the semantic understanding of SLMs. Our approach generates multiple speech continuations from a given prompt and uses semantic metrics to create preference data for Direct Preference Optimization (DPO). We evaluate the framework using ZeroSpeech 2021 benchmarks for lexical and syntactic modeling, the spoken version of the StoryCloze dataset for semantic coherence, and other speech generation metrics, including the GPT4-o score and human evaluation. Experimental results show that our method achieves state-of-the-art performance for SLMs on most benchmarks, highlighting the importance of preference optimization to improve the semantics of SLMs.
format Preprint
id arxiv_https___arxiv_org_abs_2411_01834
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Align-SLM: Textless Spoken Language Models with Reinforcement Learning from AI Feedback
Lin, Guan-Ting
Shivakumar, Prashanth Gurunath
Gourav, Aditya
Gu, Yile
Gandhe, Ankur
Lee, Hung-yi
Bulyko, Ivan
Computation and Language
Audio and Speech Processing
While textless Spoken Language Models (SLMs) have shown potential in end-to-end speech-to-speech modeling, they still lag behind text-based Large Language Models (LLMs) in terms of semantic coherence and relevance. This work introduces the Align-SLM framework, which leverages preference optimization inspired by Reinforcement Learning with AI Feedback (RLAIF) to enhance the semantic understanding of SLMs. Our approach generates multiple speech continuations from a given prompt and uses semantic metrics to create preference data for Direct Preference Optimization (DPO). We evaluate the framework using ZeroSpeech 2021 benchmarks for lexical and syntactic modeling, the spoken version of the StoryCloze dataset for semantic coherence, and other speech generation metrics, including the GPT4-o score and human evaluation. Experimental results show that our method achieves state-of-the-art performance for SLMs on most benchmarks, highlighting the importance of preference optimization to improve the semantics of SLMs.
title Align-SLM: Textless Spoken Language Models with Reinforcement Learning from AI Feedback
topic Computation and Language
Audio and Speech Processing
url https://arxiv.org/abs/2411.01834