Knowledge Commons of Institute of Automation,CAS
Learning with Average Top-k Loss | |
Fan, Yanbo1,3,4; Lyu, Siwei1; Ying, Yiming2; Hu, Bao-Gang3,4 | |
2017 | |
会议名称 | Neural Information Processing Systems (NIPS) |
会议日期 | 2017 |
会议地点 | Long Beach, CA, USA |
摘要 |
In this work, we introduce the average top-k (ATk) loss as a new aggregate loss for supervised learning, which is the average over the k largest individual losses over a training dataset. We show that the ATk loss is a natural generalization of the two widely used aggregate losses, namely the average loss and the maximum loss, but can combine their advantages and mitigate their drawbacks to better adapt to different data distributions. Furthermore, it remains a convex function over all individual losses, which can lead to convex optimization problems that can be solved effectively with conventional gradient-based methods. We provide an intuitive interpretation of the ATk loss based on its equivalent effect on the continuous individual loss functions, suggesting that it can reduce the penalty on correctly classified data. We further give a learning theory analysis of MATk learning on the classification calibration of the ATk loss and the error bounds of ATk-SVM. We demonstrate the applicability of minimum average top-k learning for binary classification and regression using synthetic and real datasets. |
关键词 | Supervised Learning Aggregate Loss Average Top-k |
收录类别 | EI |
文献类型 | 会议论文 |
条目标识符 | http://ir.ia.ac.cn/handle/173211/19993 |
专题 | 多模态人工智能系统全国重点实验室_多媒体计算 |
通讯作者 | Lyu, Siwei |
作者单位 | 1.Department of Computer Science, University at Albany, SUNY 2.Department of Mathematics and Statistics, University at Albany, SUNY 3.National Laboratory of Pattern Recognition, CASIA 4.University of Chinese Academy of Sciences |
第一作者单位 | 中国科学院自动化研究所 |
推荐引用方式 GB/T 7714 | Fan, Yanbo,Lyu, Siwei,Ying, Yiming,et al. Learning with Average Top-k Loss[C],2017. |
条目包含的文件 | 下载所有文件 | |||||
文件名称/大小 | 文献类型 | 版本类型 | 开放类型 | 使用许可 | ||
Learning with Averag(532KB) | 会议论文 | 开放获取 | CC BY-NC-SA | 浏览 下载 |
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。
修改评论