计算机视觉预训练数据的应用特殊性

Gabriel Peters, Scott Couwenhoven, Derek Walvoord, Carl Salvaggio
{"title":"计算机视觉预训练数据的应用特殊性","authors":"Gabriel Peters, Scott Couwenhoven, Derek Walvoord, Carl Salvaggio","doi":"10.1117/12.3013088","DOIUrl":null,"url":null,"abstract":"In an era of immense data generation, unlocking the full potential of Machine Learning (ML) hinges on overcoming the limitations posed by the scarcity of labeled data. In Computer Vision (CV) research, algorithm design must consider this shift and focus instead on the abundance of unlabeled imagery. In recent years, there has been a notable trend within the community toward Self-Supervised Learning (SSL) methods that can leverage this untapped data pool. ML practice promotes self-supervised pre-training for generalized feature extraction on a diverse unlabeled dataset followed by supervised transfer learning on a smaller set of labeled, application-specific images. This shift in learning methods elicits conversation about the importance of pre-training data composition for optimizing downstream performance. We evaluate models with varying measures of similarity between pre-training and transfer learning data compositions. Our findings indicate that front-end embeddings sufficiently generalize learned image features independent of data composition, leaving transfer learning to inject the majority of application-specific understanding into the model. Composition may be irrelevant in self-supervised pre-training, suggesting target data is a primary driver of application specificity. Thus, pre-training deep learning models with application-specific data, which is often difficult to acquire, is not necessary for reaching competitive downstream performance. The capability to pre-train on more accessible datasets invites more flexibility in practical deep learning.","PeriodicalId":178341,"journal":{"name":"Defense + Commercial Sensing","volume":"3 13","pages":"1305803 - 1305803-14"},"PeriodicalIF":0.0000,"publicationDate":"2024-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Application specificity of data for pre-training in computer vision\",\"authors\":\"Gabriel Peters, Scott Couwenhoven, Derek Walvoord, Carl Salvaggio\",\"doi\":\"10.1117/12.3013088\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"In an era of immense data generation, unlocking the full potential of Machine Learning (ML) hinges on overcoming the limitations posed by the scarcity of labeled data. In Computer Vision (CV) research, algorithm design must consider this shift and focus instead on the abundance of unlabeled imagery. In recent years, there has been a notable trend within the community toward Self-Supervised Learning (SSL) methods that can leverage this untapped data pool. ML practice promotes self-supervised pre-training for generalized feature extraction on a diverse unlabeled dataset followed by supervised transfer learning on a smaller set of labeled, application-specific images. This shift in learning methods elicits conversation about the importance of pre-training data composition for optimizing downstream performance. We evaluate models with varying measures of similarity between pre-training and transfer learning data compositions. Our findings indicate that front-end embeddings sufficiently generalize learned image features independent of data composition, leaving transfer learning to inject the majority of application-specific understanding into the model. Composition may be irrelevant in self-supervised pre-training, suggesting target data is a primary driver of application specificity. Thus, pre-training deep learning models with application-specific data, which is often difficult to acquire, is not necessary for reaching competitive downstream performance. The capability to pre-train on more accessible datasets invites more flexibility in practical deep learning.\",\"PeriodicalId\":178341,\"journal\":{\"name\":\"Defense + Commercial Sensing\",\"volume\":\"3 13\",\"pages\":\"1305803 - 1305803-14\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-06-06\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Defense + Commercial Sensing\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1117/12.3013088\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Defense + Commercial Sensing","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1117/12.3013088","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0

摘要

在一个产生大量数据的时代,要充分释放机器学习(ML)的潜力,关键在于克服标记数据稀缺所带来的限制。在计算机视觉(CV)研究中,算法设计必须考虑到这一转变,转而关注大量无标记图像。近年来,业界出现了一种明显的趋势,即采用自监督学习(SSL)方法来利用这一尚未开发的数据池。ML 实践提倡在一个多样化的未标记数据集上进行自我监督预训练,以进行通用特征提取,然后在一个较小的已标记的特定应用图像集上进行监督迁移学习。学习方法的这种转变引起了关于预训练数据组成对优化下游性能的重要性的讨论。我们评估了预训练和迁移学习数据组成之间不同相似度的模型。我们的研究结果表明,前端嵌入能充分泛化所学图像特征,而不受数据组成的影响,因此迁移学习能为模型注入大部分特定应用的理解。在自我监督的预训练中,组成可能无关紧要,这表明目标数据是应用特异性的主要驱动力。因此,用特定应用数据对深度学习模型进行预训练(通常很难获取),对于达到有竞争力的下游性能并非必要。在更容易获取的数据集上进行预训练的能力为实际深度学习带来了更大的灵活性。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
查看原文
分享 分享
微信好友 朋友圈 QQ好友 复制链接
本刊更多论文
Application specificity of data for pre-training in computer vision
In an era of immense data generation, unlocking the full potential of Machine Learning (ML) hinges on overcoming the limitations posed by the scarcity of labeled data. In Computer Vision (CV) research, algorithm design must consider this shift and focus instead on the abundance of unlabeled imagery. In recent years, there has been a notable trend within the community toward Self-Supervised Learning (SSL) methods that can leverage this untapped data pool. ML practice promotes self-supervised pre-training for generalized feature extraction on a diverse unlabeled dataset followed by supervised transfer learning on a smaller set of labeled, application-specific images. This shift in learning methods elicits conversation about the importance of pre-training data composition for optimizing downstream performance. We evaluate models with varying measures of similarity between pre-training and transfer learning data compositions. Our findings indicate that front-end embeddings sufficiently generalize learned image features independent of data composition, leaving transfer learning to inject the majority of application-specific understanding into the model. Composition may be irrelevant in self-supervised pre-training, suggesting target data is a primary driver of application specificity. Thus, pre-training deep learning models with application-specific data, which is often difficult to acquire, is not necessary for reaching competitive downstream performance. The capability to pre-train on more accessible datasets invites more flexibility in practical deep learning.
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
自引率
0.00%
发文量
0
期刊最新文献
Enhanced robot state estimation using physics-informed neural networks and multimodal proprioceptive data Exploring MOF-based micromotors as SERS sensors Adaptive object detection algorithms for resource constrained autonomous robotic systems Adaptive SIF-EKF estimation for fault detection in attitude control experiments A homogeneous low-resolution face recognition method using correlation features at the edge
×
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
现在去查看 取消
×
提示
确定
0
微信
客服QQ
Book学术公众号 扫码关注我们
反馈
×
意见反馈
请填写您的意见或建议
请填写您的手机或邮箱
已复制链接
已复制链接
快去分享给好友吧!
我知道了
×
扫码分享
扫码分享
Book学术官方微信
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术
文献互助 智能选刊 最新文献 互助须知 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 Book学术 All rights reserved.
ghs 京公网安备 11010802042870号 京ICP备2023020795号-1