基于轻量级图卷积和隐式反馈增强的多样化推荐

doi:10.11896/jsjkx.230900038

摘要/Abstract

摘要： 近年来,研究人员一直在努力提高推荐系统的准确性,而忽视了多样化对用户满意度的重要影响。目前大多数多样化推荐算法在传统算法生成的准确性候选列表后施加多样性约束进行后处理。然而,这种解耦设计总是导致推荐系统的次优状态。与此同时,尽管利用图卷积神经(Graph Convolution Networks,GCN)的推荐算法在提高推荐准确性方面的有效性已得到证实,但用于推荐的适用性和多样性设计仍然被忽视。此外,推荐算法采用用户购买这一单一的显式反馈无可避免地陷入“推荐过剩”。因此,提出一种端到端的多样化轻量级图卷积网络推荐模型(DiversifiedLight Graph Convolution Networks Recommendation,DLGCRec)来克服以上弊端。首先,将图卷积简化为轻量级图卷积(Light Graph Convolution Networks,LGCN)以便于推荐,并利用轻量级图卷积将多样化推向上游准确性匹配推荐过程。然后,在轻量级图卷积的采样阶段,利用引入了用户隐式反馈的多样性增强负采样来探索用户的多样化偏好。最后,利用多层特征融合策略捕获节点的完整特征嵌入,提升推荐性能。在真实数据集上进行实验,结果验证了DLGCRec在适用推荐和提升多样性方面的有效性。进一步的消融研究证实,DLGCRec有效地缓解了准确性-多样性困境。

Abstract: In recent years,researchers have been striving to improve the accuracy of recommendation systems while ignoring the critical impact of diversity on user satisfaction.Most current diversifiedrecommendation algorithms impose diversity constraints after the accuracy candidate list generated by traditional post-processing algorithms.However,this decoupled design consistently results in a sub-optimal system.Meanwhile,although the effectiveness of recommendation algorithms using graph convolution networks(GCN) in improving recommendation accuracy has been demonstrated,the applicability and diversity design for recommendation remain neglected.In addition,recommendation algorithms employing a single explicit user feedback of purchasing inevitably fall into “recommendation overload”.Therefore,an end-to-end diversified light graph convolution networks recommendation(DLGCRec) is proposed to overcome these drawbacks.Firstly,GCN is simplified to light graph convolution networks(LGCN) to be suitable for recommendation,and LGCN is utilized to push diversity upstream to the recommendation process of accuracy match.Then,in the sampling phase of LGCN,diversity-boosted negative sampling that introduces user implicit feedback is utilized to explore the user’s diversified preferences.Finally,a multi-layer feature fusion strategy is utilized to capture the complete feature embedding of the nodes to enhance the recommendation performance.Experimental results on real datasets validate the effectiveness of DLGCRec in applying in recommendations and enhancing diversity.Further ablation studies confirm that DLGCRec effectively mitigates the accuracy-diversity dilemma.

Key words: Recommendation systems, Diversity, Graph convolution networks, Implicit feedback, Accuracy-diversity dilemma

中图分类号:

TP391

黄春淦, 王桂平, 吴波, 白鑫. 基于轻量级图卷积和隐式反馈增强的多样化推荐[J]. 计算机科学, 2024, 51(6A): 230900038-11. https://doi.org/10.11896/jsjkx.230900038

HUANG Chungan, WANG Guiping, WU Bo, BAI Xin. Diversified Recommendation Based on Light Graph Convolution Networks and ImplicitFeedback Enhancement[J]. Computer Science, 2024, 51(6A): 230900038-11. https://doi.org/10.11896/jsjkx.230900038

参考文献

[1]WU L,HE X,WANG X,et al.A Survey on Accuracy-oriented Neural Recommendation:From Collaborative Filtering to Information-rich Recommendation[C]//IEEE Transactions on Knowledge and Data Engineering.2022.
[2]HU R,PU P.Helping Users Perceive Recommendation Diversity[C]//DiveRS@RecSys.Chicago,USA:ACM,2011:43-50.
[3]JIANG Z,LIU H,FU B,et al.Recommendation in heterogeneous information networks based on generalized random walk model and bayesian personalized ranking[C]//Proceedings of the Eleventh ACM International Conference on Web Search and Data Mining.Marina Del Rey CA USA:ACM,2018:288-296.
[4]PAUDEL B,BERNSTEIN A.Random Walks with Erasure:Di-versifying Personalized Recommendations on Social and Information Networks[C]//Proceedings of the Web Conference 2021.Ljubljana Slovenia:ACM,2021:2046-2057.
[5]HE X,DENG K,WWANG X,et al.LightGCN:Simplifying and Power-ing Graph Convolution Network for Recommendation[M].arXiv,2020.
[6]MAO K,ZHU J,XIAO X,et al.UltraGCN:Ultra Simplification of Graph Convolutional Networks for Recommendation[M].arXiv,2021.
[7]WANG H,ZHAO M,XIE X,et al.Knowledge graph convolu-tional networks for recommender systems[C]//The World Wide Web Conference on WWW ’19.San Francisco,CA,USA:ACM Press,2019:3307-3313.
[8]WANG X,HE X,WANG M,et al.Neural Graph Collaborative Filtering[C]//Proceedings of the 42nd International ACM SIGIR Conference on Research and Development in Information Retrieval.2019:165-174.
[9]ZHENG Y,GAO C,CHEN L,et al.DGCN:Diversified Rec-ommendation with Graph Convolutional Networks[C]//Procee-dings of the Web Conference 2021.New York,NY,USA:Asso-ciation for Computing Machinery,2021:401-412.
[10]LIANG Y,QIAN T,LI Q,et al.Enhancing Domain-Level and User-Level Adaptivity in Diversified Recommendation[C]//Proceedings of the 44th International ACM SIGIR Conference on Research and Development in Information Retrieval.New York,NY,USA:Association for Computing Machinery,2021:747-756.
[11]LIU Y,XIAO Y,WU Q,et al.Diversified Interactive Recom-mendation with Implicit Feedback[C]//Proceedings of the AAAI Conference on Artificial Intelligence.2020:4932-4939.
[12]ZHOU T,KUSCSIK Z,LIU J G,et al.Solving the apparent diversity-accuracy dilemma of recommender systems[J].Procee-dings of the National Academy of Sciences of the United States of America,2010,107(10):4511-4515.
[13]SHA C,WU X,NIU J.A framework for recommending relevant and diverse items[C]//Proceedings of the Twenty-Fifth International Joint Conference on Artificial Intelligence.New York,USA:AAAI,2016:3868-3874.
[14]MEI D,HUANG N,LI X.Light Graph Convolutional Collaborative Filtering With Multi-Aspect Information[J].IEEE Access,2021,9:34433-34441.
[15]CHEN L,ZHANG G,ZHOU E.Fast Greedy MAP Inference for Determinantal Point Process to Improve Recommendation Diversi-ty[C]//Advances in Neural Information Processing Systems:31.Curran Associates,Inc.,2018.
[16]DEREZIN＇SKI M.Fast determinantal point processes via distor-tion-free intermediate sampling[M].arXiv,2019.
[17]GAN L,NURBAKOVA D,LAPORTE L,et al.Enhancing Recommendation Diversity using Determinantal Point Processes on Knowledge Graphs[C]//Proceedings of the 43rd International Acm Sigir Conference on Research and Development in Information Retrieval(sigir ’20).New York:Assoc Computing Machi-nery,2020:2001-2004.
[18]ABDOLLAHPOURI H,BURKE R,MOBASHER B.Managing popularity bias in recommender systems with personalized re-ranking[J].arXiv:1901.07555,2019.
[19]CARBONELL J,GOLDSTEIN J.The use of MMR,diversity-based reranking for reordering documents and producing summaries[C]//Proceedings of the 21st Annual International ACM SIGIR Conference on Research and Development in Information Retrieval(SIGIR ’98).Melbourne,Australia:ACM Press,1998:335-336.
[20]CHENG P,WANG S,MA J,et al.Learning to Recommend Accurate and Diverse Items[C]//Proceedings of the 26th International Conference on World Wide Web.Perth Australia:International World Wide Web Conferences Steering Committee,2017:183-192.
[21]ASHKAN A,KVETON B,BERKOVSKY S,et al.OptimalGreedy Diversity for Recommendation[C]//Twenty-Fourth International Joint Conference on Artificial Intelligence.2015.
[22]ZUO Y,LIU S,ZHOU Y.DTGCF:Diversified Tag-Aware Re-commendation with Graph Collaborative Filtering[J].Applied Sciences,2023,13(5):2945.
[23]YANG L,WANG S,TAO Y,et al.DGRec:Graph Neural Net-work for Recommendation with Diversified Embedding Genera-tion[C]//Proceedings of the Sixteenth ACM International Conference on Web Search and Data Mining.2023:661-669.
[24]ABADAL S,JAIN A,GUIRADO R,et al.Computing GraphNeural Networks:A Survey from Algorithms to Accelerators[J].ACM Computing Surveys,2022,54(9):1-38.
[25]WAIKHOM L,PATGIRI R.Graph Neural Networks:Me-thods,Applications,and Opportunities[M].arXiv,2021.
[26]ZHOU Y,ZHENG H,HUANG X,et al.Graph Neural Net-works:Taxonomy,Advances and Trends[J].ACM Transactions on Intelligent Systems and Technology,2022,13(1):1-54.
[27]KIPF T N,WELLING M.Semi-supervised classification withgraph convolutional networks[J].arXiv:1609.02907,2016.
[28]LI Z,LIU Z,HUANG J,et al.MV-GCN:Multi-View GraphConvolutional Networks for Link Prediction[J].IEEE Access,2019,7:176317-176328.
[29]GAO X,FENG F,HUANG H,et al.Food recommendation with graph convolutional network[J].Information Sciences,2022,584:170-183.
[30]KIM J Y,CHO S B.A systematic analysis and guidelines of graph neural networks for practical applications[J].Expert Systems with Applications,2021,184:115466.
[31]LI Y.A graph convolution network based on improved density clustering for recommendation system[J].Information Techno-logy and Control,2022,51(1):18-31.
[32]TANG X,YANG J,XIONG D,et al.Knowledge-enhanced graph convolutional network for recommendation[J].Multimedia Tools and Applications,2022,81(20):28899-28916.
[33]CHEN L,WU L,HONG R,et al.Revisiting Graph Based Col-laborative Filtering:A Linear Residual Graph Convolutional Network Approach[J].Proceedings of the AAAI Conference on Artificial Intelligence,2020,34(1):27-34.
[34]WU F,ZHANG T,SOUZA Jr.A H de,et al.Simplifying Graph Convolutional Networks[M].arXiv,2019.
[35]MIKOLOV T,SUTSKEVER I,CHEN K,et al.Distributed Re-presentations of Words and Phrases and their Compositionality[M].arXiv,2013.
[36]YANG Z,DING M,ZHOU C,et al.Understanding NegativeSampling in Graph Representation Learning[C]//Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining.Virtual Event CA USA:ACM,2020:1666-1676.
[37]ZHU Y,XU Y,LIU Q,et al.An Empirical Study of Graph Con-trastive Learning[M].arXiv,2021.
[38]ZHU Y,XU Y,YU F,et al.Graph Contrastive Learning with Adaptive Augmentation[C]//Proceedings of the Web Conference 2021.2021:2069-2080.
[39]WEI Y,WANG X,NIE L,et al.Graph-Refined Convolutional Network for Multimedia Recommendation with Implicit Feed-back[C]//Proceedings of the 28th ACM International Confe-rence on Multimedia.New York,NY,USA:Association for Computing Machinery,2020:3541-3549.
[40]ADOMAVICIUS G,KWON Y.Improving Aggregate Recom-mendation Diversity Using Ranking-Based Techniques[J].IEEE Transactions on Knowledge and Data Engineering,2012,24(5):896-911.
[41]REDDI S J,KALE S,KUMAR S.On the Convergence of Adam and Beyond[M].arXiv,2019.
[42]WILHELM M,RAMANATHAN A,BONOMO A,et al.Practi-cal Diversified Recommendations on YouTube with Determinantal Point Processes[C]//Proceedings of the 27th ACM International Conference on Information and Knowledge Management.Torino Italy:ACM,2018:2165-2173.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed