TY - JOUR
T1 - Complex heterogeneity learning
T2 - A theoretical and empirical study
AU - Yang, Pei
AU - Tan, Qi
AU - He, Jingrui
N1 - Funding Information:
This work is supported by National Natural Science Foundation of China ( No. 61976092 and 61473123 ), Natural Science Foundation of Guangdong (No. 2017A030313370 and 2018A030313356 ), National Science Foundation ( No. IIS-1552654, IIS-1813464, IIS-1651203, and CNS-1629888 ), U.S.Department of Homeland Security ( No. 17STQAC00001-02-00 ), and an IBM Faculty Award. The views are those of the authors and should not be interpreted as representing the official policies of the funding agencies or the governments.
Publisher Copyright:
© 2020 Elsevier Ltd
PY - 2020/11
Y1 - 2020/11
N2 - Data heterogeneity such as task heterogeneity, view heterogeneity, and instance heterogeneity often co-exist in many real-world applications including insider threat detection, traffic prediction, brain image analysis, quality control in manufacturing processes, etc. However, most of the existing techniques might not take fully advantage of the rich heterogeneity. To address this problem, we propose a novel graph-based approach named M3 to simultaneously model triple heterogeneity in a principled framework. The main idea is to employ the hybrid graphs to jointly model the task relatedness, view consistency, and bag-instance correlation by enhancing the labeling consistency between nearby nodes on the graphs. Furthermore, we analyze the generalization performance of the proposed method based on Rademacher complexity, which sheds light on the benefits of jointly modeling multiple types of heterogeneity. The resulting optimization problem is challenging since the objective function is non-smooth and non-convex. We propose an iterative algorithm based on block coordinate descent and bundle method to solve the problem. Experimental results on various datasets demonstrate the effectiveness of the proposed method.
AB - Data heterogeneity such as task heterogeneity, view heterogeneity, and instance heterogeneity often co-exist in many real-world applications including insider threat detection, traffic prediction, brain image analysis, quality control in manufacturing processes, etc. However, most of the existing techniques might not take fully advantage of the rich heterogeneity. To address this problem, we propose a novel graph-based approach named M3 to simultaneously model triple heterogeneity in a principled framework. The main idea is to employ the hybrid graphs to jointly model the task relatedness, view consistency, and bag-instance correlation by enhancing the labeling consistency between nearby nodes on the graphs. Furthermore, we analyze the generalization performance of the proposed method based on Rademacher complexity, which sheds light on the benefits of jointly modeling multiple types of heterogeneity. The resulting optimization problem is challenging since the objective function is non-smooth and non-convex. We propose an iterative algorithm based on block coordinate descent and bundle method to solve the problem. Experimental results on various datasets demonstrate the effectiveness of the proposed method.
KW - Heterogeneous learning
KW - Multi-instance learning
KW - Multi-task learning
KW - Multi-view learning
UR - http://www.scopus.com/inward/record.url?scp=85087403183&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85087403183&partnerID=8YFLogxK
U2 - 10.1016/j.patcog.2020.107519
DO - 10.1016/j.patcog.2020.107519
M3 - Article
AN - SCOPUS:85087403183
SN - 0031-3203
VL - 107
JO - Pattern Recognition
JF - Pattern Recognition
M1 - 107519
ER -