面向服务质量感知云API推荐系统的数据投毒攻击检测方法

doi:10.11959/j.issn.1000-436x.2023161

通信学报 ›› 2023, Vol. 44 ›› Issue (8): 155-167.doi: 10.11959/j.issn.1000-436x.2023161

• 学术论文 • 上一篇

面向服务质量感知云API推荐系统的数据投毒攻击检测方法

陈真¹^,², 乞文超¹, 鲍泰宇¹, 申利民¹^,²

¹ 燕山大学信息科学与工程学院，河北秦皇岛 066004
² 燕山大学河北省计算机虚拟技术与系统集成重点实验室，河北秦皇岛 066004

修回日期:2023-07-27 出版日期:2023-08-01 发布日期:2023-08-01
作者简介:陈真（1987- ），男，陕西宝鸡人，博士，燕山大学副教授、博士生导师，主要研究方向为服务计算、推荐系统和服务化软件开发等
乞文超（1998- ），女，河北邢台人，燕山大学硕士生，主要研究方向为服务计算、云API安全和推荐算法
鲍泰宇（1999- ），男，河北石家庄人，燕山大学硕士生，主要研究方向为服务计算、协同推荐算法和数据投毒攻击与检测等
申利民（1962- ），男，黑龙江佳木斯人，博士，燕山大学教授、博士生导师，主要研究方向为协同计算、服务计算和信息安全等
基金资助:
国家自然科学基金资助项目(62102348);国家自然科学基金资助项目(62276226);河北省自然科学基金资助项目(F2022203012);中央引导地方科技发展资金资助项目(236Z0103G);河北省教育厅高等学校科技计划基金资助项目(QN2020183);河北省创新能力提升计划基金资助项目(22567626H)

Data poisoning attack detection approach for quality of service aware cloud API recommender system

Zhen CHEN¹^,², Wenchao QI¹, Taiyu BAO¹, Limin SHEN¹^,²

¹ School of Information Science and Engineering, Yanshan University, Qinhuangdao 066004, China
² Hebei Key Laboratory of Computer Virtual Technology and System Integration, Yanshan University, Qinhuangdao 066004, China

Revised:2023-07-27 Online:2023-08-01 Published:2023-08-01
Supported by:
The National Natural Science Foundation of China(62102348);The National Natural Science Foundation of China(62276226);The Natural Science Foundation of Hebei Province(F2022203012);The Central Guiding Local Science and Technology Development Fund Project(236Z0103G);The Science and Technology Research Project of Hebei University(QN2020183);The Innovation Capability Improvement Plan Project of Hebei Province(22567626H)

摘要/Abstract

摘要：

针对现有研究通常假设云API推荐系统的服务质量数据是可靠的，忽略了开放网络环境中恶意用户对云API推荐系统的数据投毒攻击的问题，提出了一种基于多特征融合的数据投毒攻击检测方法。首先，依据设计的相似性度量函数构建用户连通网络图，并利用 Node2vec 捕获用户邻域特征；其次，采用稀疏自编码器挖掘用户服务质量深度特征，并构建基于服务质量数据加权平均偏差的用户解释特征。进一步，融合用户邻域特征、服务质量深度特征和解释特征建立基于支持向量机的虚假用户检测模型，并使用网格搜索和交替迭代优化策略学习模型参数，继而实现虚假用户检测。最后，通过多组实验验证了所提方法的有效性和优越性，实现了服务质量感知云API推荐系统在数据端的投毒攻击防御。

关键词: 推荐系统, 云API, 服务质量, 数据投毒, 攻击检测

Abstract:

To solve the problem that existing studies usually assumed that the QoS data of cloud API recommender system was reliable, ignoring the data poisoning attack on cloud API recommender system by malicious users in open network environment, a data poisoning attack detection approach based on multi-feature fusion was proposed.Firstly, a user connected network graph was constructed based on the designed similarity function, and users’ neighborhood features were captured using Node2vec.Secondly, sparse auto-encoder was used to mine user QoS deep feature, and user interpretation feature based on QoS data weighted average deviation was designed.Furthermore, a fake user detection model based on support vector machine was established by integrating user neighborhood feature, QoS deep feature, and interpretation feature, the model parameters were learned using grid search and alternating iterative optimization strategy to complete fake user detection.Finally, the effectiveness and superiority of the proposed approach were verified through extensive experiments, realizing the poison attack defense against QoS aware cloud API recommender system at the data side.

Key words: recommender system, cloud API, quality of service, data poisoning, attack detection

中图分类号:

TP311

陈真, 乞文超, 鲍泰宇, 申利民. 面向服务质量感知云API推荐系统的数据投毒攻击检测方法[J]. 通信学报, 2023, 44(8): 155-167.

Zhen CHEN, Wenchao QI, Taiyu BAO, Limin SHEN. Data poisoning attack detection approach for quality of service aware cloud API recommender system[J]. Journal on Communications, 2023, 44(8): 155-167.

图/表 11

表1

图1

图2

图3

表2

图4

图5

图6

表3

在响应时间数据集上消融变体实验对比结果"

方法		随机攻击			潮流攻击			均值攻击
方法	准确率	召回率	F1值	准确率	召回率	F1值	准确率	召回率	F1值
NQI-N	0.565	0.500	0.529	0.617	0.667	0.635	0.577	0.594	0.580
NQI-S	0.841	0.817	0.829	0.927	0.951	0.941	0.902	0.947	0.927
NQI-W	0.972	0.951	0.972	0.941	0.918	0.928	0.958	0.922	0.948
NQI-NS	0.554	0.482	0.512	0.524	0.584	0.552	0.590	0.549	0.552
NQI-NW	0.594	0.569	0.573	0.561	0.538	0.545	0.561	0.527	0.548
NQI-SW	0.958	0.945	0.950	0.932	0.910	0.924	0.945	0.962	0.955
NQI	$0 . 989$	$0 . 964$	$0 . 955$	$0 . 975$	$0 . 959$	$0 . 968$	$0 . 970$	$0 . 982$	$0 . 979$

表3

表4

在吞吐量数据集上消融变体实验对比结果"

方法		随机攻击			潮流攻击			均值攻击
方法	准确率	召回率	F1值	准确率	召回率	F1值	准确率	召回率	F1值
NQI-N	0.732	0.754	0.742	0.695	0.752	0.617	0.715	0.744	0.729
NQI-S	0.941	0.964	0.955	0.849	0.869	0.854	0.874	0.832	0.857
NQI-W	0.964	0.970	0.966	0.957	0.922	0.941	0.958	0.967	0.962
NQI-NS	0.533	0.630	0.574	0.585	0.562	0.542	0.523	0.583	0.554
NQI-NW	0.583	0.659	0.611	0.570	0.551	0.561	0.614	0.655	0.639
NQI-SW	0.961	0.957	0.958	0.953	0.948	0.955	0.938	0.920	0.928
NQI	$0 . 978$	$0 . 985$	$0 . 980$	$0 . 964$	$0 . 982$	$0 . 975$	$0 . 972$	$0 . 955$	$0 . 961$

表4

图7

参考文献 24

[1]	GONG W W , ZHANG X Y , CHEN Y F ,et al. DAWAR:diversity-aware Web APIs recommendation for mashup creation based on correlation graph[C]// Proceedings of the 45th International ACM SIGIR Conference on Research and Development in Information Retrieval. New York:ACM Press, 2022: 395-404.
[2]	QI L Y , LIN W M , ZHANG X Y ,et al. A correlation graph based approach for personalized and compatible Web APIs recommendation in mobile APP development[J]. IEEE Transactions on Knowledge and Data Engineering, 2023,35(6): 5444-5457.
[3]	CHEN Z , PAN M S , HE P F ,et al. Context and auto-interaction are all you need:towards context embedding based QoS prediction via automatic feature interaction for high quality cloud API delivery[J]. Future Generation Computer Systems, 2022,128: 265-281.
[4]	SHU Y J , ZHANG J H , ZHANG W E ,et al. IQSrec:an efficient and diversified skyline services recommendation on incomplete QoS[J]. IEEE Transactions on Services Computing, 2023,16(3): 1934-1948.
[5]	CAO B , LIU J , WEN Y ,et al. QoS-aware service recommendation based on relational topic model and factorization machines for IoT mashup applications[J]. Journal of Parallel and Distributed Computing, 2019,132: 177-189.
[6]	陈真, 乞文超, 贺鹏飞 ,等. 云应用程序编程接口安全研究综述：威胁与防护[J]. 电子与信息学报, 2023,45(1): 371-382.
	CHEN Z , QI W C , HE P F ,et al. A survey for cloud application programming interface security:threats and protection[J]. Journal of Electronics ＆ Information Technology, 2023,45(1): 371-382.
[7]	ZHENG Z B , LI X L , TANG M D ,et al. Web service QoS prediction via collaborative filtering:a survey[J]. IEEE Transactions on Services Computing, 2022,15(4): 2455-2472.
[8]	YE F H , LIN Z W , CHEN C ,et al. Outlier-resilient Web service QoS prediction[C]// Proceedings of the Web Conference 2021. New York:ACM Press, 2021: 3099-3110.
[9]	MANIKRAO U S , PRABHAKAR T V . Dynamic selection of Web services with recommendation system[C]// Proceedings of International Conference on Next Generation Web Services Practices. Piscataway:IEEE Press, 2005: 117-121.
[10]	RAN S P . A model for Web services discovery with QoS[J]. ACM SIGecom Exchanges, 2003,4(1): 1-10.
[11]	YANG J , CHEN Y , RAHARDJA S . Neighborhood representative for improving outlier detectors[J]. Information Sciences, 2023,625: 192-205.
[12]	BURRA V B , PACHALA S . An improved proxy re-encryption scheme using resource optimization and authentication protocol[J]. International Journal of System Assurance Engineering and Management, 2023:doi.org/10.1007/s13198-022-01809-9.
[13]	GUPTA P , YADAV K , GUPTA B B ,et al. A novel data poisoning attack in federated learning based on inverted loss function[J]. Computers ＆ Security, 2023,130: 1-8.
[14]	ZHANG H X , WANG D Y , ZHANG W ,et al. QoS prediction in intelligent edge computing based on feature learning[J]. Journal of Cloud Computing, 2023,12(1): 1-16.
[15]	LI J H , WU H , CHEN J P ,et al. Topology-aware neural model for highly accurate QoS prediction[J]. IEEE Transactions on Parallel and Distributed Systems, 2022,33(7): 1538-1552.
[16]	ZHU J , LI B , WANG J ,et al. BGCL:bi-subgraph network based on graph contrastive learning for cold-start QoS prediction[J]. Knowledge-Based Systems, 2023,263: 1-11.
[17]	HA J , PARK S . NCMD:Node2vec-based neural collaborative filtering for predicting MiRNA-disease association[J]. IEEE/ACM Transactions on Computational Biology and Bioinformatics, 2022,20(2): 1257-1268.
[18]	XIA D W , ZHENG Y L , BAI Y ,et al. A parallel grid-search-based SVM optimization algorithm on Spark for passenger hotspot prediction[J]. Multimedia Tools and Applications, 2022,81(19): 27523-27549.
[19]	YANG Y T , ZHENG Z B , NIU X D ,et al. A location-based factorization machine model for Web service QoS prediction[J]. IEEE Transactions on Services Computing, 2021,14(5): 1264-1277.
[20]	MEHTA B , NEJDL W . Unsupervised strategies for shilling detection and robust collaborative filtering[J]. User Modeling and User-Adapted Interaction, 2009,19(1): 65-97.
[21]	YANG F , GAO M , YU J L ,et al. Detection of shilling attack based on Bayesian model and user embedding[C]// Proceedings of 2018 IEEE 30th International Conference on Tools with Artificial Intelligence. Piscataway:IEEE Press, 2018: 639-646.
[22]	ZHOU Q , WU J , DUAN L . Recommendation attack detection based on deep learning[J]. Journal of Information Security and Applications, 2020,52: 1-13.
[23]	郝耀军, 张付志 . 基于深度自动编码器的托攻击集成检测方法[J]. 计算机工程与应用, 2019,55(1): 9-22.
	HAO Y , ZHANG F . Ensemble detection method for shilling attacks based on deep sparse autoencoder[J]. Computer Engineering and Applications, 2019,55(1): 9-22.
[24]	MAATEN L , HINTON G . Visualizing data using t-SNE[J]. Journal of Machine Learning Research, 2008,9: 2579-2605.

攻击模型	A_S	A_F	A_Ф	A_T
随机攻击	∅	r_random	∅	r_max/min
潮流攻击	流行的云API	r_random	∅	r_max
均值攻击	∅	r_average	∅	r_max/min

数据	值
响应时间范围/s	(0,20]
平均响应时间/s	0.908
响应时间数据集用户数量/个	339
响应时间数据集云API数量/个	5 818
吞吐量范围/(kbit·s^-1)	(0,1 000]
平均吞吐量/(kbit·s^-1)	47.561
吞吐量数据集用户数量/个	339
吞吐量数据集云API数量/个	5 801

面向服务质量感知云API推荐系统的数据投毒攻击检测方法

Data poisoning attack detection approach for quality of service aware cloud API recommender system

在线阅读

PDF下载

可视化

摘要/Abstract

引用本文

使用本文

图/表 11

参考文献 24

相关文章 15

Metrics

推荐阅读 0

[1]	巩小雪, 庞嘉豪, 张琦涵, 徐长乐, 秦文帅, 郭磊. 基于机器学习的光网络干扰攻击检测、识别与恢复方法[J]. 通信学报, 2023, 44(7): 159-170.
[2]	陈真, 陈文辉, 刘啸威, 尤殿龙, 刘林林, 申利民. 功能互补关系增强的云API推荐方法[J]. 通信学报, 2023, 44(6): 125-137.
[3]	李元诚, 秦永泰. 基于深度强化学习的软件定义安全中台QoS实时优化算法[J]. 通信学报, 2023, 44(5): 181-192.
[4]	荆有波, 曹清越, 朱瑞. MgdFlow：微电网场景下的多粒度数据流管理算法[J]. 通信学报, 2023, 44(10): 94-102.
[5]	刘建勋, 丁领航, 康国胜, 曹步清, 肖勇. 基于特征深度融合的Web服务QoS联合预测[J]. 通信学报, 2022, 43(7): 215-226.
[6]	张达敏, 王义, 邹诚诚, 赵沛雯, 张琳娜. 认知异构蜂窝网络中改进蜉蝣算法的资源分配策略[J]. 通信学报, 2022, 43(6): 156-167.
[7]	谢丽霞, 李雪鸥, 杨宏宇, 张良, 成翔. 基于样本特征强化的APT攻击多阶段检测方法[J]. 通信学报, 2022, 43(12): 66-76.
[8]	段雪源, 付钰, 王坤, 李彬. 基于简单统计特征的LDoS攻击检测方法[J]. 通信学报, 2022, 43(11): 53-64.
[9]	李俊娥, 陆秋余, 刘剑, 袁凯, 田薇, 彭弼君. 智能变电站通信业务优先级及其队列调度方法[J]. 通信学报, 2021, 42(7): 25-40.
[10]	汪硕, 黄玉栋, 黄韬, 霍如, 刘韵洁. 基于软件定义的时间敏感网络跨域调度机制[J]. 通信学报, 2021, 42(10): 1-9.
[11]	卢毅,徐梦颖,周杰. 基于改进的免疫克隆蛙跳算法的多约束QoS路由优化研究[J]. 通信学报, 2020, 41(5): 141-149.
[12]	孙伟,张鹏,何永全,邢丽超. 内网环境下基于时空事件关联的攻击检测方法[J]. 通信学报, 2020, 41(1): 33-41.
[13]	李罡,吴志军. 基于多QoS约束条件的广域信息管理系统任务调度算法[J]. 通信学报, 2019, 40(7): 27-37.
[14]	黄韬,汪硕,黄玉栋,郑尧,刘江,刘韵洁. 确定性网络研究综述[J]. 通信学报, 2019, 40(6): 160-176.
[15]	苏命峰,王国军,李仁发. 基于利益相关视角的多维QoS云资源调度方法[J]. 通信学报, 2019, 40(6): 102-115.