从全局到可靠性的两阶段早期退出

IF 7.3 2区 计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE
Jianing He, Qi Zhang, Hongyun Zhang, Duoqian Miao
{"title":"从全局到可靠性的两阶段早期退出","authors":"Jianing He,&nbsp;Qi Zhang,&nbsp;Hongyun Zhang,&nbsp;Duoqian Miao","doi":"10.1049/cit2.70010","DOIUrl":null,"url":null,"abstract":"<p>Early exiting has shown significant potential in accelerating the inference of pre-trained language models (PLMs) by allowing easy samples to exit from shallow layers. However, existing early exiting methods primarily rely on local information from individual samples to estimate prediction uncertainty for making exiting decisions, overlooking the global information provided by the sample population. This impacts the estimation of prediction uncertainty, compromising the reliability of exiting decisions. To remedy this, inspired by principal component analysis (PCA), the authors define a residual score to capture the deviation of features from the principal space of the sample population, providing a global perspective for estimating prediction uncertainty. Building on this, a two-stage exiting strategy is proposed that integrates global information from residual scores with local information from energy scores at both the decision and feature levels. This strategy incorporates three-way decisions to enable more reliable exiting decisions for boundary region samples by delaying judgement. Extensive experiments on the GLUE benchmark validate that the method achieves an average speed-up ratio of 2.17× across all tasks with minimal performance degradation. Additionally, it surpasses the state-of-the-art E-LANG by <span></span><math>\n <semantics>\n <mrow>\n <mn>11</mn>\n <mi>%</mi>\n </mrow>\n <annotation> $11\\%$</annotation>\n </semantics></math> in model acceleration, along with a performance improvement of 0.6 points, demonstrating a better performance-efficiency trade-off.</p>","PeriodicalId":46211,"journal":{"name":"CAAI Transactions on Intelligence Technology","volume":"10 4","pages":"1019-1032"},"PeriodicalIF":7.3000,"publicationDate":"2025-04-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ietresearch.onlinelibrary.wiley.com/doi/epdf/10.1049/cit2.70010","citationCount":"0","resultStr":"{\"title\":\"Two-Stage Early Exiting From Globality Towards Reliability\",\"authors\":\"Jianing He,&nbsp;Qi Zhang,&nbsp;Hongyun Zhang,&nbsp;Duoqian Miao\",\"doi\":\"10.1049/cit2.70010\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p>Early exiting has shown significant potential in accelerating the inference of pre-trained language models (PLMs) by allowing easy samples to exit from shallow layers. However, existing early exiting methods primarily rely on local information from individual samples to estimate prediction uncertainty for making exiting decisions, overlooking the global information provided by the sample population. This impacts the estimation of prediction uncertainty, compromising the reliability of exiting decisions. To remedy this, inspired by principal component analysis (PCA), the authors define a residual score to capture the deviation of features from the principal space of the sample population, providing a global perspective for estimating prediction uncertainty. Building on this, a two-stage exiting strategy is proposed that integrates global information from residual scores with local information from energy scores at both the decision and feature levels. This strategy incorporates three-way decisions to enable more reliable exiting decisions for boundary region samples by delaying judgement. Extensive experiments on the GLUE benchmark validate that the method achieves an average speed-up ratio of 2.17× across all tasks with minimal performance degradation. Additionally, it surpasses the state-of-the-art E-LANG by <span></span><math>\\n <semantics>\\n <mrow>\\n <mn>11</mn>\\n <mi>%</mi>\\n </mrow>\\n <annotation> $11\\\\%$</annotation>\\n </semantics></math> in model acceleration, along with a performance improvement of 0.6 points, demonstrating a better performance-efficiency trade-off.</p>\",\"PeriodicalId\":46211,\"journal\":{\"name\":\"CAAI Transactions on Intelligence Technology\",\"volume\":\"10 4\",\"pages\":\"1019-1032\"},\"PeriodicalIF\":7.3000,\"publicationDate\":\"2025-04-11\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"https://ietresearch.onlinelibrary.wiley.com/doi/epdf/10.1049/cit2.70010\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"CAAI Transactions on Intelligence Technology\",\"FirstCategoryId\":\"94\",\"ListUrlMain\":\"https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/cit2.70010\",\"RegionNum\":2,\"RegionCategory\":\"计算机科学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"CAAI Transactions on Intelligence Technology","FirstCategoryId":"94","ListUrlMain":"https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/cit2.70010","RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE","Score":null,"Total":0}
引用次数: 0

摘要

通过允许简单的样本从浅层退出,早期退出在加速预训练语言模型(PLMs)的推理方面显示出巨大的潜力。然而,现有的早期退出方法主要依靠个体样本的局部信息来估计退出决策的预测不确定性,忽略了样本总体提供的全局信息。这影响了预测不确定性的估计,损害了现有决策的可靠性。为了解决这个问题,受主成分分析(PCA)的启发,作者定义了残差分数来捕捉样本总体主空间的特征偏差,为估计预测不确定性提供了一个全局视角。在此基础上,提出了一种两阶段退出策略,该策略将残差评分的全局信息与决策和特征级别的能量评分的局部信息相结合。该策略采用三向决策,通过延迟判断,使边界区域样本的退出决策更加可靠。在GLUE基准测试上的大量实验验证了该方法在所有任务中实现了2.17倍的平均加速比,并且性能下降最小。此外,它在模型加速方面超过了最先进的E-LANG 11%,性能提高了0.6分,表现出更好的性能效率权衡。
本文章由计算机程序翻译,如有差异,请以英文原文为准。

Two-Stage Early Exiting From Globality Towards Reliability

Two-Stage Early Exiting From Globality Towards Reliability

Two-Stage Early Exiting From Globality Towards Reliability

Two-Stage Early Exiting From Globality Towards Reliability

Early exiting has shown significant potential in accelerating the inference of pre-trained language models (PLMs) by allowing easy samples to exit from shallow layers. However, existing early exiting methods primarily rely on local information from individual samples to estimate prediction uncertainty for making exiting decisions, overlooking the global information provided by the sample population. This impacts the estimation of prediction uncertainty, compromising the reliability of exiting decisions. To remedy this, inspired by principal component analysis (PCA), the authors define a residual score to capture the deviation of features from the principal space of the sample population, providing a global perspective for estimating prediction uncertainty. Building on this, a two-stage exiting strategy is proposed that integrates global information from residual scores with local information from energy scores at both the decision and feature levels. This strategy incorporates three-way decisions to enable more reliable exiting decisions for boundary region samples by delaying judgement. Extensive experiments on the GLUE benchmark validate that the method achieves an average speed-up ratio of 2.17× across all tasks with minimal performance degradation. Additionally, it surpasses the state-of-the-art E-LANG by 11 % $11\%$ in model acceleration, along with a performance improvement of 0.6 points, demonstrating a better performance-efficiency trade-off.

求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
来源期刊
CAAI Transactions on Intelligence Technology
CAAI Transactions on Intelligence Technology COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE-
CiteScore
11.00
自引率
3.90%
发文量
134
审稿时长
35 weeks
期刊介绍: CAAI Transactions on Intelligence Technology is a leading venue for original research on the theoretical and experimental aspects of artificial intelligence technology. We are a fully open access journal co-published by the Institution of Engineering and Technology (IET) and the Chinese Association for Artificial Intelligence (CAAI) providing research which is openly accessible to read and share worldwide.
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:604180095
Book学术官方微信