GSM: Global Semantic Memory
文献类型：期刊论文
| 作者 | Li, Jiahao1,2,3,4; Chen, Yiqiang1,2,3; Xing, Yunbing1,3,4; Gu, Yang1,2,3; Lan, Xiangyuan2 |
| 刊名 | PATTERN RECOGNITION
 |
| 出版日期 | 2026 |
| 卷号 | 169页码:16 |
| 关键词 | Anomaly detection Unsupervised learning Memory |
| ISSN号 | 0031-3203 |
| DOI | 10.1016/j.patcog.2025.111950 |
| 英文摘要 | Unsupervised anomaly detection is to detect previously unseen rare samples without any prior knowledge about them. With the emergence of deep learning, many methods employ normal data reconstruction to train detection models, which is expected to yield relatively large errors when reconstructing anomalies. However, recent studies find that anomalies can be overgeneralized, resulting in reconstruction errors as small as normal samples. In this paper, we examine the anomaly overgeneralization problem and propose global semantic information learning. Normal and anomalous samples may share the same local feature such as textures, edges, and corners, but have separability at the global semantic level. To address this, we propose the global semantic memory with novel cascade architecture designed to capture global semantic information in the latent space and introduce a configurable sparsification and random forgetting mechanism. Our proposed method achieves state-of-the-art experimental results on different public benchmarks, without the introduction of any additional auxiliary loss. |
| 资助项目 | Innovation Funding of ICT, CAS[E461020] ; Hunan Provincial Natural Science Foundation of China[2023JJ70009] ; Beijing Municipal Science & Technology Commission[Z221100002722009] ; Youth Innovation Promotion Association CAS[2021101] ; Science and Technology Innovation Program of Hunan Province[2022RC4006] ; Project of Pengcheng Laboratory[PCL2023A08] |
| WOS研究方向 | Computer Science ; Engineering |
| 语种 | 英语 |
| WOS记录号 | WOS:001532675700001 |
| 出版者 | ELSEVIER SCI LTD |
| 源URL | [http://119.78.100.204/handle/2XEOYT63/42076]  |
| 专题 | 中国科学院计算技术研究所期刊论文_英文 |
| 通讯作者 | Chen, Yiqiang |
| 作者单位 | 1.Chinese Acad Sci, Inst Comp Technol, Beijing Key Lab Mobile Comp & Pervas Device, Beijing 100190, Peoples R China 2.Pengcheng Lab, Shenzhen 518000, Peoples R China 3.Univ Chinese Acad Sci, Beijing 100190, Peoples R China 4.AIER Eye Hosp Grp Co Ltd, Changsha 410015, Peoples R China |
| 推荐引用方式 GB/T 7714 | Li, Jiahao,Chen, Yiqiang,Xing, Yunbing,et al. GSM: Global Semantic Memory[J]. PATTERN RECOGNITION,2026,169:16. |
| APA | Li, Jiahao,Chen, Yiqiang,Xing, Yunbing,Gu, Yang,&Lan, Xiangyuan.(2026).GSM: Global Semantic Memory.PATTERN RECOGNITION,169,16. |
| MLA | Li, Jiahao,et al."GSM: Global Semantic Memory".PATTERN RECOGNITION 169(2026):16. |
入库方式： OAI收割
来源：计算技术研究所
浏览0
下载0
收藏0
其他版本
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。