https://scholars.lib.ntu.edu.tw/handle/123456789/581326
標題: | Increasingly packing multiple facial-informatics modules in a unified deep-learning model via lifelong learning | 作者: | Hung S.C.Y Lee J.-H Wan T.S.T Chen C.-H Chan Y.-M CHU-SONG CHEN |
關鍵字: | Deep learning; Deep neural networks; Informatics; Iterative methods; Multimedia systems; Neural networks; Compact model; Continual learning; Facial Expressions; Gender identification; Learning models; Life long learning; Multi-task model; New functions; Face recognition | 公開日期: | 2019 | 起(迄)頁: | 339-343 | 來源出版物: | ICMR 2019 - Proceedings of the 2019 ACM International Conference on Multimedia Retrieval | 摘要: | Simultaneously running multiple modules is a key requirement for a smart multimedia system for facial applications including face recognition, facial expression understanding, and gender identification. To effectively integrate them, a continual learning approach to learn new tasks without forgetting is introduced. Unlike previous methods growing monotonically in size, our approach maintains the compactness in continual learning. The proposed packing-andexpanding method is effective and easy to implement, which can iteratively shrink and enlarge the model to integrate new functions. Our integrated multitask model can achieve similar accuracy with only 39.9% of the original size. ? 2019 Association for Computing Machinery. |
URI: | https://www.scopus.com/inward/record.uri?eid=2-s2.0-85068074204&doi=10.1145%2f3323873.3325053&partnerID=40&md5=6f3a83d486a435fe7d8fe4dbf19d8672 https://scholars.lib.ntu.edu.tw/handle/123456789/581326 |
DOI: | 10.1145/3323873.3325053 |
顯示於: | 資訊工程學系 |
在 IR 系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。