Saved in:
Bibliographic Details
Main Authors: Lei, Shuyu, Liu, Lingen, Yang, Jiaolong, Jiao, Yasen, Yang, Yuxiang, Yang, Yushu, Guo, Xiang
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2407.14875
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910536714682368
author Lei, Shuyu
Liu, Lingen
Yang, Jiaolong
Jiao, Yasen
Yang, Yuxiang
Yang, Yushu
Guo, Xiang
author_facet Lei, Shuyu
Liu, Lingen
Yang, Jiaolong
Jiao, Yasen
Yang, Yuxiang
Yang, Yushu
Guo, Xiang
contents Existing auto-regressive language models have demonstrated a remarkable capability to perform a new task with just a few examples in prompt, without requiring any additional training. In order to extend this capability to a multi-modal setting (i.e. speech and language), this paper introduces the Seal model, an abbreviation for speech language model. It incorporates a novel alignment method, in which Kullback-Leibler divergence loss is performed to train a projector that bridges a frozen speech encoder with a frozen language model decoder. The resulting Seal model exhibits robust performance as a few-shot learner on two speech understanding tasks. Additionally, consistency experiments are conducted to validate its robustness on different pre-trained language models.
format Preprint
id arxiv_https___arxiv_org_abs_2407_14875
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Seal: Advancing Speech Language Models to be Few-Shot Learners
Lei, Shuyu
Liu, Lingen
Yang, Jiaolong
Jiao, Yasen
Yang, Yuxiang
Yang, Yushu
Guo, Xiang
Computation and Language
Existing auto-regressive language models have demonstrated a remarkable capability to perform a new task with just a few examples in prompt, without requiring any additional training. In order to extend this capability to a multi-modal setting (i.e. speech and language), this paper introduces the Seal model, an abbreviation for speech language model. It incorporates a novel alignment method, in which Kullback-Leibler divergence loss is performed to train a projector that bridges a frozen speech encoder with a frozen language model decoder. The resulting Seal model exhibits robust performance as a few-shot learner on two speech understanding tasks. Additionally, consistency experiments are conducted to validate its robustness on different pre-trained language models.
title Seal: Advancing Speech Language Models to be Few-Shot Learners
topic Computation and Language
url https://arxiv.org/abs/2407.14875