Unlabeled data selection for active learning in image classification
Abstract Active Learning has emerged as a viable solution for addressing the challenge of labeling extensive amounts of data in data-intensive applications such as computer vision and neural machine translation. The main objective of Active Learning is to automatically identify a subset of unlabeled...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2024-01-01
|
| Series: | Scientific Reports |
| Online Access: | https://doi.org/10.1038/s41598-023-50598-z |
| _version_ | 1797363434946297856 |
|---|---|
| author | Xiongquan Li Xukang Wang Xuhesheng Chen Yao Lu Hongpeng Fu Ying Cheng Wu |
| author_facet | Xiongquan Li Xukang Wang Xuhesheng Chen Yao Lu Hongpeng Fu Ying Cheng Wu |
| author_sort | Xiongquan Li |
| collection | DOAJ |
| description | Abstract Active Learning has emerged as a viable solution for addressing the challenge of labeling extensive amounts of data in data-intensive applications such as computer vision and neural machine translation. The main objective of Active Learning is to automatically identify a subset of unlabeled data samples for annotation. This identification process is based on an acquisition function that assesses the value of each sample for model training. In the context of computer vision, image classification is a crucial task that typically requires a substantial training dataset. This research paper introduces innovative selection methods within the Active Learning framework, aiming to identify informative images from unlabeled datasets while minimizing the number of required training data. The proposed methods, namely Similari-ty-based Selection, Prediction Probability-based Selection, and Competence-based Active Learning, have been extensively evaluated through experiments conducted on popular datasets like Cifar10 and Cifar100. The experimental results demonstrate that the proposed methods outperform random selection and conventional selection techniques. The superior performance of the novel selection methods underscores their effectiveness in enhancing the Active Learning process for image classification tasks. |
| first_indexed | 2024-03-08T16:20:18Z |
| format | Article |
| id | doaj.art-4546edb5e9794cb095bcf2f3eef087b0 |
| institution | Directory Open Access Journal |
| issn | 2045-2322 |
| language | English |
| last_indexed | 2024-03-08T16:20:18Z |
| publishDate | 2024-01-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Scientific Reports |
| spelling | doaj.art-4546edb5e9794cb095bcf2f3eef087b02024-01-07T12:22:19ZengNature PortfolioScientific Reports2045-23222024-01-0114111310.1038/s41598-023-50598-zUnlabeled data selection for active learning in image classificationXiongquan Li0Xukang Wang1Xuhesheng Chen2Yao Lu3Hongpeng Fu4Ying Cheng Wu5Faculty of Information Engineering and Automation, Kunming University of Science and TechnologySage IT Consulting GroupThe University of North Carolina at Chapel HillUniversity of BristolKhoury College of Computer Sciences, Northeastern UniversityUniversity of WashingtonAbstract Active Learning has emerged as a viable solution for addressing the challenge of labeling extensive amounts of data in data-intensive applications such as computer vision and neural machine translation. The main objective of Active Learning is to automatically identify a subset of unlabeled data samples for annotation. This identification process is based on an acquisition function that assesses the value of each sample for model training. In the context of computer vision, image classification is a crucial task that typically requires a substantial training dataset. This research paper introduces innovative selection methods within the Active Learning framework, aiming to identify informative images from unlabeled datasets while minimizing the number of required training data. The proposed methods, namely Similari-ty-based Selection, Prediction Probability-based Selection, and Competence-based Active Learning, have been extensively evaluated through experiments conducted on popular datasets like Cifar10 and Cifar100. The experimental results demonstrate that the proposed methods outperform random selection and conventional selection techniques. The superior performance of the novel selection methods underscores their effectiveness in enhancing the Active Learning process for image classification tasks.https://doi.org/10.1038/s41598-023-50598-z |
| spellingShingle | Xiongquan Li Xukang Wang Xuhesheng Chen Yao Lu Hongpeng Fu Ying Cheng Wu Unlabeled data selection for active learning in image classification Scientific Reports |
| title | Unlabeled data selection for active learning in image classification |
| title_full | Unlabeled data selection for active learning in image classification |
| title_fullStr | Unlabeled data selection for active learning in image classification |
| title_full_unstemmed | Unlabeled data selection for active learning in image classification |
| title_short | Unlabeled data selection for active learning in image classification |
| title_sort | unlabeled data selection for active learning in image classification |
| url | https://doi.org/10.1038/s41598-023-50598-z |
| work_keys_str_mv | AT xiongquanli unlabeleddataselectionforactivelearninginimageclassification AT xukangwang unlabeleddataselectionforactivelearninginimageclassification AT xuheshengchen unlabeleddataselectionforactivelearninginimageclassification AT yaolu unlabeleddataselectionforactivelearninginimageclassification AT hongpengfu unlabeleddataselectionforactivelearninginimageclassification AT yingchengwu unlabeleddataselectionforactivelearninginimageclassification |