.png){kind=icon}
返回
Data Augmentation by Guided Deep Interpolation
.png){kind=icon}
PRE
AI
DOI：10.1016/j.asoc.2021.107680.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
State-of-the-art machine learning algorithms require large amount of high quality data. In practice, however, the sample size is commonly low and data is imbalanced along different class labels. Low sample size and imbalanced class distribution can significantly deteriorate the predictive performance of machine learning models. In order to overcome data quality issues, we propose a novel data augmentation method, Guided Deep Interpolation (GDI). It is based on a convolutional auto-encoder network, which is equipped with an auxiliary linear self-expressive layer. The network is trained by minimizing a composite objective function so that to extract the underlying clustered structure of semantic similarities of data points while high reconstruction quality is also preserved. The trained network is used to define a sampling strategy and a synthetic data generation procedure. Making use of the weights of the self-expressive layer, we introduce a measure of semantic variability to quantify how similar a data point to other data points on average. Based on the proposed measure of semantic variability, a joint distribution is defined. Using the distribution we can draw pairs of similar data points so that one point is semantically underrepresented (isolated) while its pair possesses relatively high semantic variability. A sampled pair is interpolated in the deep feature space of the network so that to increase semantic variability while preserve class label of the semantically underrepresented data point. The trained decoder is used to determine pixel space representations of latent space interpolations. The resulting data augmentation procedure generates synthetic samples by increasing the semantic variability of semantically underrepresented instances in a class label preserving way. Our experimental results show that the proposed method outperforms traditional and generative model-based data augmentation methods on low sample size and imbalanced data sets. (C) 2021 Elsevier B.V. All rights reserved.
Keyword:
Autoencoders
Data augmentation
Imbalanced data
Self-expressiveness
Interpolation
AI总结
对已上传原文的论文进行重点信息的提取，主要内容包括：简要概述、研究摘要、背景介绍、关键亮点、图文解析、展望与总结。
期刊
IF:
6.6
论文数:
1.4W
被引数:
4.8W
机构
引用论文
Disease activity in primary progressive multiple sclerosis: a systematic review and meta-analysis
OAAI
OAAI分享收藏Tinnitus Retraining Therapy (TRT) as a Method for Treatment of Tinnitus and Hyperacusis Patients耳鸣再训练疗法 (TRT) 作为治疗耳鸣和高亢患者的方法
PREAI
分享收藏Arsenic removal from aqueous solutions by adsorption using novel MIL-53(Fe) as a highly efficient adsorbent使用新型MIL-53(Fe) 作为高效吸附剂通过吸附从水溶液中去除砷.png)
RSC Advances
RSC AdvancesIF0
PREAI
分享收藏Structural study of lanthanides(III) in aqueous nitrate and chloride solutions by EXAFS通过EXAFS对硝酸盐和氯化物水溶液中镧系元素 (III) 的结构研究
PREAI
分享收藏
查看更多内容