Regularizing Deep Networks with Semantic Data Augmentation

计算机科学 人工智能 深度学习 上下界 交叉熵 模式识别(心理学) 算法 机器学习 正规化(语言学) 数学 数学分析
作者
Yulin Wang,Gao Huang,Shiji Song,Xuran Pan,Yitong Xia,Cheng Wu
出处
期刊:IEEE Transactions on Pattern Analysis and Machine Intelligence [IEEE Computer Society]
卷期号:: 1-1 被引量:127
标识
DOI:10.1109/tpami.2021.3052951
摘要

Data augmentation is widely known as a simple yet surprisingly effective technique for regularizing deep networks. Conventional data augmentation schemes, e.g., flipping, translation or rotation, are low-level, data-independent and class-agnostic operations, leading to limited diversity for augmented samples. To this end, we propose a novel semantic data augmentation algorithm to complement traditional approaches. The proposed method is inspired by the intriguing property that deep networks are effective in learning linearized features, i.e., certain directions in the deep feature space correspond to meaningful semantic transformations, e.g., changing the background or view angle of an object. Based on this observation, translating training samples along many such directions in the feature space can effectively augment the dataset for more diversity. To implement this idea, we first introduce a sampling based method to obtain semantically meaningful directions efficiently. Then, an upper bound of the expected cross-entropy (CE) loss on the augmented training set is derived by assuming the number of augmented samples goes to infinity, yielding a highly efficient algorithm. In fact, we show that the proposed implicit semantic data augmentation (ISDA) algorithm amounts to minimizing a novel robust CE loss, which adds minimal extra computational cost to a normal training procedure. In addition to supervised learning, ISDA can be applied to semi-supervised learning tasks under the consistency regularization framework, where ISDA amounts to minimizing the upper bound of the expected KL-divergence between the augmented features and the original features. Although being simple, ISDA consistently improves the generalization performance of popular deep models (e.g., ResNets and DenseNets) on a variety of datasets, i.e., CIFAR-10, CIFAR-100, SVHN, ImageNet, and Cityscapes. Code for reproducing our results is available at https://github.com/blackfeather-wang/ISDA-for-Deep-Networks.
最长约 10秒,即可获得该文献文件

科研通智能强力驱动
Strongly Powered by AbleSci AI
科研通是完全免费的文献互助平台,具备全网最快的应助速度,最高的求助完成率。 对每一个文献求助,科研通都将尽心尽力,给求助人一个满意的交代。
实时播报
77完成签到,获得积分10
2秒前
眼睛大的小熊猫完成签到,获得积分10
2秒前
清爽难破完成签到 ,获得积分10
4秒前
4秒前
NingCp完成签到,获得积分10
5秒前
5秒前
6秒前
6秒前
6秒前
传奇3的应助被quantu采纳,获得10
6秒前
踏实以松发布了新的文献求助10
6秒前
丘比特的应助被幽凡采纳,获得10
8秒前
林林完成签到 ,获得积分10
9秒前
坚强孤容发布了新的文献求助10
11秒前
11秒前
12秒前
迷你的寄云完成签到,获得积分10
14秒前
14秒前
15秒前
李麟发布了新的文献求助10
15秒前
思源的应助被邵蹦蹦采纳,获得10
16秒前
17秒前
xuejingling发布了新的文献求助10
17秒前
molihuakai的应助被长情豁采纳,获得10
17秒前
科研通AI6.2的应助被123采纳,获得10
17秒前
哈哈客发布了新的文献求助10
19秒前
在水一方的应助被百里采纳,获得10
19秒前
搜集达人的应助被坚强孤容采纳,获得10
20秒前
21秒前
21秒前
Hello的应助被徐XU采纳,获得10
22秒前
七七完成签到 ,获得积分10
22秒前
rongjian完成签到 ,获得积分10
23秒前
人间慈悲的应助被辛勤的西牛采纳,获得10
23秒前
24秒前
24秒前
张二发布了新的文献求助10
26秒前
26秒前
surge发布了新的文献求助10
26秒前
27秒前
高分求助中
(应助此贴封号)【重要!!请各用户(尤其是新用户)详细阅读】【科研通的精品贴汇总】 10000
拉伸高分子熔体中缠结行为的模拟和小角中子散射研究 1700
The World Atlas of Wine 9th edition 1000
奈特基础组织学:配套组织病理学(第4版)(Netter's Essential Histology With Correlated Histopathology) 800
八旗制度の研究 600
高分子基功能界面水蒸发材料与应用研究 500
外场增强的水汽发电研究 500
热门求助领域 (近24小时)
化学 材料科学 医学 生物 纳米技术 计算机科学 工程类 有机化学 化学工程 生物化学 物理 复合材料 内科学 催化作用 光电子学 人工智能 细胞生物学 无机化学 物理化学 数学
热门帖子
关注 科研通微信公众号,转发送积分 8186734
求助须知:如何正确求助?哪些是违规求助? 9575007
关于积分的说明 21839114
捐赠科研通 7784289
什么是DOI,文献DOI怎么找? 3434808
关于科研通互助平台的介绍 2559206
邀请新用户注册赠送积分活动 2481611