arXivDaily arXiv每日学术速递 周一至周五更新
北美退伍军人节(Veterans Day),12 日无更新
arXiv 2409.08199cs.CLcs.AIcs.SDeess.AS
收藏

AudioBERT:音频知识增强语言模型

AudioBERT: Audio Knowledge Augmented Language Model

POSTECH(浦项科技大学);HJ AILAB(HJ人工智能实验室);Inha University(仁荷大学)

Hyunjong Ok, Suho Yoo, Jaeho Lee

AI总结:

针对纯文本预训练语言模型听觉知识匮乏的问题,提出基于检索的AudioBERT方法,通过检测提示听觉知识跨度、注入音频知识并结合低秩适配增强BERT,在自建AuditoryBench基准上取得优异性能。

摘要:

近期研究发现,在纯文本数据集上预训练的语言模型往往缺乏基础的视觉知识,例如日常物品的颜色。受这一观察启发,我们探究这类模型在听觉知识方面是否存在类似缺陷。为回答该问题,我们构建了名为AuditoryBench的新数据集,包含两项用于评估听觉知识的全新任务。基于该基准的分析结果,我们发现语言模型同样存在听觉知识严重匮乏的问题。为解决这一局限,我们提出AudioBERT,这是一种通过基于检索的方法增强BERT听觉知识的新方法。首先,我们检测提示中的听觉知识跨度,以高效查询我们的检索模型。随后,我们将音频知识注入BERT,并在需要音频知识时启用低秩适配以实现有效适配。实验表明,AudioBERT效果显著,在AuditoryBench上取得了优异性能。数据集与代码可访问https://github.com/HJ-Ok/AudioBERT获取。

英文摘要:

Recent studies have identified that language models, pretrained on text-only datasets, often lack elementary visual knowledge, \textit{e.g.,} colors of everyday objects. Motivated by this observation, we ask whether a similar shortcoming exists in terms of the \textit{auditory} knowledge. To answer this question, we construct a new dataset called AuditoryBench, which consists of two novel tasks for evaluating auditory knowledge. Based on our analysis using the benchmark, we find that language models also suffer from a severe lack of auditory knowledge. To address this limitation, we propose AudioBERT, a novel method to augment the auditory knowledge of BERT through a retrieval-based approach. First, we detect auditory knowledge spans in prompts to query our retrieval model efficiently. Then, we inject audio knowledge into BERT and switch on low-rank adaptation for effective adaptation when audio knowledge is required. Our experiments demonstrate that AudioBERT is quite effective, achieving superior performance on the AuditoryBench. The dataset and code are available at \bulurl{https://github.com/HJ-Ok/AudioBERT}.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑