基于自适应多任务卷积神经网络的脑网络分类方法

邢新颖; 冀俊忠; 姚垚

doi:10.7544/issn1000-1239.2020.20190186

基于自适应多任务卷积神经网络的脑网络分类方法

(北京工业大学信息学部计算机学院北京 100124) (xinying.xing@emails.bjut.edu.cn)

基金项目: 国家自然科学基金项目(61672065)

详细信息

中图分类号: TP18
计量
- 文章访问数: 1254
- HTML全文浏览量: 7
- PDF下载量: 334
出版历程
- 发布日期: 2020-06-30

Brain Networks Classification Based on an Adaptive Multi-Task Convolutional Neural Networks

(College of Computer, Faculty of Information Technology, Beijing University of Technology, Beijing 100124)

Funds: This work was supported by the National Natural Science Foundation of China (61672065).

摘要

摘要: 脑网络分类是脑科学研究中的一项重要课题.近年来，基于卷积神经网络的脑网络分类方法已经成为一个前沿热点.然而，目前仍难以对数据维度高、样本量小的脑网络数据进行精准分类.由于不同人群的临床表型与其脑网络差异存在着一定的依存关系，极有可能为脑网络分类提供辅助信息，故提出一种新的基于自适应多任务卷积神经网络的脑网络分类方法.该方法引入临床表型预测作为辅助任务，通过多任务卷积神经网络的共享表示机制来为脑网络分类提供有用信息；同时为了降低实验成本和人工操作带来的误差，提出了一种新的自适应方法来代替人工调整多任务学习中各个子任务的权重.在ABIDE I(autism brain imaging data exchange I)数据集上的实验结果表明：引入临床表型预测任务的多任务卷积神经网络能够获得更好的脑网络分类结果，而且自适应多任务学习方法能够进一步提升脑网络的分类性能.
- 脑网络分类 /
- 卷积神经网络 /
- 多任务学习 /
- 临床表型 /
- 自适应方法
Abstract: Brain networks classification is an important subject in brain science. In recent years, brain networks classification based on convolutional neural networks has become a hot topic. However, it is still difficult to accurately classify brain network data with high dimension and small sample size. Due to the close relationship between different clinical phenotypes and brain networks of different populations, it is highly possible to provide auxiliary information for the brain networks classification. Therefore, we propose a new brain networks classification method based on an adaptive multi-task convolutional neural network in this paper. Firstly, the clinical phenotype predictions are introduced as different auxiliary tasks and the shared representation mechanism of multi-task convolutional neural networks is used to provide general and useful information for brain networks classification. Then, in order to reduce the experimental cost and the error caused by the manual operation, a new adaptive method is proposed to substitute for manual adjustments of the weight of every task in the multi-task learning. The experimental results on the autism brain imaging data exchange I (ABIDE I) dataset show that the multi-task convolutional neural networks which introduce clinical phenotype predictions can achieve better classification results. Moreover, the adaptive multi-task learning method can further improve the performance of brain networks classification.
- brain networks classification /
- convolutional neural network /
- multi-task learning /
- clinical phenotype /
- adaptive method

HTML全文

参考文献(0)

施引文献(15)

期刊类型引用(9)

1.	李振华，王泓懿，李洋，林灏，杨昕磊. 大规模复杂终端网络的云原生强化设计. 计算机研究与发展. 2024(01): 2-19 . 本站查看
2.	赵旭康，刘晓锋，徐洁. 融合多样频度与分布差异的Android恶意软件检测. 计算机工程与设计. 2024(02): 390-395 . 百度学术
3.	方加娟，丁乙恒. 基于关联规则的Android恶意软件检测技术. 电脑与信息技术. 2024(03): 115-118 . 百度学术
4.	陈志强，韩萌，武红鑫，李慕航，张喜龙. 分段加权的概念漂移检测方法. 计算机应用. 2023(03): 776-784 . 百度学术
5.	李汇来，杨斌，于秀丽，唐晓梅. 软件缺陷预测模型可解释性对比. 计算机科学. 2023(05): 21-30 . 百度学术
6.	潘建文，崔展齐，林高毅，陈翔，郑丽伟. Android恶意应用的静态检测方法综述. 计算机研究与发展. 2023(08): 1875-1894 . 本站查看
7.	殷建艳. 面向云数据库的Android应用风险评估方法. 信息与电脑(理论版). 2023(17): 177-179 . 百度学术
8.	张皓. 基于深度学习的恶意软件动态检测方法研究. 电子技术与软件工程. 2022(03): 43-46 . 百度学术
9.	刘光源. 基于DoI-RNNs模型的恶意软件动态检测方法. 信息与电脑(理论版). 2022(23): 38-40 . 百度学术