吉林大学学报（理学版）2024，Vol.62Issue(1)：122-131,10.DOI:10.13413/j.cnki.jdxblxb.2023058
自动语音识别模型压缩算法综述
Compression Algorithms for Automatic Speech Recognition Models:A Survey
摘要
Abstract
With the development of deep learning technology,the number of parameters in automatic speech recognition task models was becoming increasingly large,which gradually increased the computing overhead,storage requirements and power consumption of the models,and it was difficult to deploy on resource-constrained devices.Therefore,it was of great value to compress the automatic speech recognition models based on deep learning to reduce the size of the modes while maintaining the original performance as much as possible.Aiming at the above problems,a comprehensive survey was conducted on the main works in this field in recent years,which was summarized as several methods,including knowledge distillation,model quantization,low-rank decomposition,network pruning,parameter sharing and combination models,and conducted a systematic review to provide alternative solutions for the deployment of models on resource-constrained devices.关键词
语音识别/模型压缩/知识蒸馏/模型量化/低秩分解/网络剪枝/参数共享Key words
speech recognition/model compression/knowledge distillation/model quantization/low-rank decomposition/network pruning/parameter sharing分类
信息技术与安全科学引用本文复制引用
时小虎,袁宇平,吕贵林,常志勇,邹元君..自动语音识别模型压缩算法综述[J].吉林大学学报（理学版）,2024,62(1):122-131,10.基金项目
国家自然科学基金(批准号:62272192)、吉林省科技发展计划项目(批准号:20210201080GX)、吉林省发改委项目(批准号:2021C044-1)和吉林省教育厅科研基金(批准号:JJKH20200871KJ). （批准号:62272192）
