Enhancing Robustness of Deep Networks Against Noisy Labels Based on A Two-Phase Formulation of Their Learning Behavior
文献类型:会议论文
作者 | Luo YR(罗曜儒) |
出版日期 | 2023 |
会议日期 | 2023-7 |
会议地点 | 澳大利亚-布里斯班 |
英文摘要 | In this study we propose an explicit formulation of the learning behavior of deep neural networks (DNNs) trained with noisy labels in image classification. Specifically, we show theoretically and experimentally that the training process can be divided into two phases: a learning phase in which the outputs of DNNs converge to a hidden noisy label distribution; and a memorization phase in which DNNs start to overfit until the output for each sample converges to its corresponding noisy label. This two-phase formulation enables us to resolve a common pitfall of existing methods for robust training against noisy labels based on the small-loss assumption, namely clean samples have smaller losses than noisy samples in the early training phase. We show that these methods fail when the noise transition matrix is not column diagonally maximal and that this pitfall can be fixed by a simple modification of the small-loss assumption. |
源URL | [http://ir.ia.ac.cn/handle/173211/54524] |
专题 | 模式识别国家重点实验室_计算生物学与机器智能 多模态人工智能系统全国重点实验室 |
作者单位 | 1.University of Chinese Academy of Sciences 2.Institute of Automation Chinese Academy of Science |
推荐引用方式 GB/T 7714 | Luo YR. Enhancing Robustness of Deep Networks Against Noisy Labels Based on A Two-Phase Formulation of Their Learning Behavior[C]. 见:. 澳大利亚-布里斯班. 2023-7. |
入库方式: OAI收割
来源:自动化研究所
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。