AI helps you reading Science

AI generates interpretation videos

AI extracts and analyses the key points of the paper to generate videos automatically


pub
Go Generating

AI Traceability

AI parses the academic lineage of this thesis


Master Reading Tree
Generate MRT

AI Insight

AI extracts a summary of this paper


Weibo:
We introduce a novel explainable interaction-driven user modeling algorithm to better capture the users’ interaction-level dynamic preferences in an explainable way in Sequential Recommendation tasks

Explainable Interaction-driven User Modeling over Knowledge Graph for Sequential Recommendation

Proceedings of the 27th ACM International Conference on Multimedia, pp.548-556, (2019)

Cited by: 14|Views71
EI

Abstract

Compared with the traditional recommendation system, sequential recommendation holds the ability of capturing the evolution of users' dynamic interests. Many previous studies in sequential recommendation focus on the accuracy of predicting the next item that a user might interact with, while generally ignore providing explanations why the...More

Code:

Data:

0
Introduction
  • Sequential Recommendation (SR) aims to meet the current needs of user according to her/his historical behavior sequence [17].
  • Recurrent Neural Network (RNN) is a classical algorithm in SR task, which is able to capture temporal dependencies by encoding user’s historical behaviors into a latent vector.
  • Self-attention based sequential recommendation algorithms have attracted increasing attention due to the flexibility and efficiency of the model [12, 33].
  • Those NN-based methods achieve high accuracy in recommendation task.
  • Most of those methods do not consider providing users with credible explanations while recommending
Highlights
  • Sequential Recommendation (SR) aims to meet the current needs of user according to her/his historical behavior sequence [17]
  • In order to address the problem of introducing semantic paths into SR system for capturing user’s dynamic preferences and providing accurate explainable recommendations, we propose an Explainable Interaction-driven User Modeling (EIUM) approach for SR task
  • KTUP[3] is a knowledge-enhanced translation-based user preference model. It transfers the relation embeddings as well as entity embeddings learned from Knowledge Graph (KG) to user preference model, and simultaneously training two different tasks
  • We introduce a novel explainable interaction-driven user modeling algorithm to better capture the users’ interaction-level dynamic preferences in an explainable way in SR tasks
  • The user-item interactions are constructed by several semantic paths extracted from knowledge graph, which endows the SR system the ability of accuracy and explainability
  • Experimental results show that compared with the state-of-the-art methods, our approach improves the accuracy of recommendations, and holds the explanation capacity
  • We will continue to extend EIUM in better incorporating user’s profiles and contextual information with external KG for potential preferences seeking in dealing with cold-start recommendation problems
Methods
  • BPR Bi-LSTM Bi-LSTM+att.
  • ATRank CKE KTUP EIUM AUC.
  • 0.9065 (-0.18%) 0.8800 (-3.09%) 0.8897 (-2.03%) 0.8724 (-3.93%) 0.9054 (-0.30%) 0.9187∗ (+1.17%) MAP.
Results
  • 5.3.2 Training Details.
  • The whole model is trained in an end-toend way with Adam optimizer.
  • The authors apply a grid search for the learning rate and find lr = 0.01 is the best.
  • A grid search in {2n, n = 5to9} is applied to find out the best setting of the embedding dimension, and set dim.
  • The batch size is set to 32.
  • A detailed analysis is illustrated in Section 5.6.
  • The parameters are set as suggested by the original papers
Conclusion
  • The authors introduce a novel explainable interaction-driven user modeling algorithm to better capture the users’ interaction-level dynamic preferences in an explainable way in SR tasks.
  • The user-item interactions are constructed by several semantic paths extracted from knowledge graph, which endows the SR system the ability of accuracy and explainability.
  • Extensive experiments demonstrate the superior ability of the proposed EIUM model on providing effective and convincing recommendations to users.
  • The authors will continue to extend EIUM in better incorporating user’s profiles and contextual information with external KG for potential preferences seeking in dealing with cold-start recommendation problems
Summary
  • Introduction:

    Sequential Recommendation (SR) aims to meet the current needs of user according to her/his historical behavior sequence [17].
  • Recurrent Neural Network (RNN) is a classical algorithm in SR task, which is able to capture temporal dependencies by encoding user’s historical behaviors into a latent vector.
  • Self-attention based sequential recommendation algorithms have attracted increasing attention due to the flexibility and efficiency of the model [12, 33].
  • Those NN-based methods achieve high accuracy in recommendation task.
  • Most of those methods do not consider providing users with credible explanations while recommending
  • Objectives:

    The authors aim to mine the high-level interactive representations between users and each item, instead of low-level item-based onefold representation.
  • Methods:

    BPR Bi-LSTM Bi-LSTM+att.
  • ATRank CKE KTUP EIUM AUC.
  • 0.9065 (-0.18%) 0.8800 (-3.09%) 0.8897 (-2.03%) 0.8724 (-3.93%) 0.9054 (-0.30%) 0.9187∗ (+1.17%) MAP.
  • Results:

    5.3.2 Training Details.
  • The whole model is trained in an end-toend way with Adam optimizer.
  • The authors apply a grid search for the learning rate and find lr = 0.01 is the best.
  • A grid search in {2n, n = 5to9} is applied to find out the best setting of the embedding dimension, and set dim.
  • The batch size is set to 32.
  • A detailed analysis is illustrated in Section 5.6.
  • The parameters are set as suggested by the original papers
  • Conclusion:

    The authors introduce a novel explainable interaction-driven user modeling algorithm to better capture the users’ interaction-level dynamic preferences in an explainable way in SR tasks.
  • The user-item interactions are constructed by several semantic paths extracted from knowledge graph, which endows the SR system the ability of accuracy and explainability.
  • Extensive experiments demonstrate the superior ability of the proposed EIUM model on providing effective and convincing recommendations to users.
  • The authors will continue to extend EIUM in better incorporating user’s profiles and contextual information with external KG for potential preferences seeking in dealing with cold-start recommendation problems
Tables
  • Table1: Statistics of our dataset
  • Table2: The next-one recommendation performance of all the methods across all the evaluation metrics. The best performance is boldfaced; the highest score in baseline is labeled with ‘∗’; the percentage in parentheses represents the relative improvements that baselines achieve w.r.t EIUM
Download tables as Excel
Related work
  • Sequential Recommendation (SR). Sequence modeling methods for SR mainly belong to Markov Models [4, 17] and RNNs[8, 9, 15, 27]. The scalable sequential models usually rely on Markov Chain (MC) to capture sequential patterns. L-order MC makes recommendations based on L previous actions. It cannot be directly applied to sequence-aware recommendation in most cases since data sparsity quickly leads to poor estimates of the transition matrices, and it also faces the challenge of the choice of the order [14]. Therefore, many studies are devoted to improving the MC-based approach. Inspired by the great power of Matrix Factorization (MF), Factorized Personalized Markov Chain (FPMC) combines the power of MF and MC to factorize the transition matrix over underlying MC to model personalized sequential behaviors for the problem of next-item recommendation given the last-N interactions of the user [4, 17]. RNN computes the current hidden state from the current input in the sequence and the hidden state outputted by the previous time step. The recurrent feedback mechanism memorizes the influence of each past data sample in the hidden state. It therefore makes RNN and its variants such as LSTM and GRU be able to model the temporal information for user behaviors in recommendation task [8, 9, 15, 27]. Though it is an effective way to encode users’ behavior sequences, it still suffers from several difficulties, such as hard to parallelize, time-consuming, hard to preserve long-term dependencies. The emergence of the Transformer [22] architecture tackles the problem of sequence transduction. Some studies abandon the complex and time-consuming RNN structures, and instead construct the sequence model based on self-attention mechanism and apply it to SR system. ATRank [33] takes the lead in using self-attention structure for the SR and achieves encouraging results.
Funding
  • This work was supported in part by the National Key Research and Development Program of China (No 2017YFB1002804), the National Natural Science Foundation of China under Grants 61432019, 61720106006, 61572503, 61802405, 61872424, 61702509 and 61832002, the Key Research Program of Frontier Sciences, CAS, Grant NO
Reference
  • Dzmitry Bahdanau, Kyunghyun Cho, and Yoshua Bengio. 2014. Neural machine translation by jointly learning to align and translate. arXiv preprint arXiv:1409.0473 (2014).
    Findings
  • Antoine Bordes, Nicolas Usunier, Alberto Garcia-Duran, Jason Weston, and Oksana Yakhnenko. 2013. Translating embeddings for modeling multi-relational data. In Advances in neural information processing systems. 2787–2795.
    Google ScholarFindings
  • Yixin Cao, Xiang Wang, Xiangnan He, Tat-Seng Chua, et al. 2019. Unifying Knowledge Graph Learning and Recommendation: Towards a Better Understanding of User Preferences. arXiv preprint arXiv:1902.06236 (2019).
    Findings
  • Chen Cheng, Haiqin Yang, Michael R Lyu, and Irwin King. 2013. Where You Like to Go Next: Successive Point-of-Interest Recommendation.. In IJCAI, Vol. 13. 2605–2611.
    Google ScholarLocate open access versionFindings
  • Li Gao, Hong Yang, Jia Wu, Chuan Zhou, Weixue Lu, and Yue Hu. 2018. Recommendation with multi-source heterogeneous information. structure 1, w3 (2018), w4.
    Google ScholarLocate open access versionFindings
  • Xiangnan He, Lizi Liao, Hanwang Zhang, Liqiang Nie, Xia Hu, and Tat-Seng Chua. 2017. Neural collaborative filtering. In Proceedings of the 26th International Conference on World Wide Web. International World Wide Web Conferences Steering Committee, 173–182.
    Google ScholarLocate open access versionFindings
  • Benjamin Heitmann and Conor Hayes. 2010. Using linked data to build open, collaborative recommender systems. In 2010 AAAI Spring Symposium Series.
    Google ScholarFindings
  • Balázs Hidasi, Alexandros Karatzoglou, Linas Baltrunas, and Domonkos Tikk. 2015. Session-based recommendations with recurrent neural networks. arXiv preprint arXiv:1511.06939 (2015).
    Findings
  • Balázs Hidasi, Massimo Quadrana, Alexandros Karatzoglou, and Domonkos Tikk. 2016. Parallel recurrent neural network architectures for feature-rich session-based recommendations. In Proceedings of the 10th ACM Conference on Recommender Systems. ACM, 241–248.
    Google ScholarLocate open access versionFindings
  • Binbin Hu, Chuan Shi, Wayne Xin Zhao, and Philip S Yu. 2018. Leveraging metapath based context for top-n recommendation with a neural co-attention model. In Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. ACM, 1531–1540.
    Google ScholarLocate open access versionFindings
  • Jin Huang, Wayne Xin Zhao, Hongjian Dou, Ji-Rong Wen, and Edward Y Chang. 2018. Improving sequential recommendation with knowledge-enhanced memory networks. In The 41st International ACM SIGIR Conference on Research & Development in Information Retrieval. ACM, 505–514.
    Google ScholarLocate open access versionFindings
  • Xiaowen Huang, Shengsheng Qian, Quan Fang, Jitao Sang, and Changsheng Xu. 2018. CSAN: Contextual Self-Attention Network for User Sequential Recommendation. In 2018 ACM Multimedia Conference on Multimedia Conference. ACM, 447–455.
    Google ScholarLocate open access versionFindings
  • Tomas Mikolov, Edouard Grave, Piotr Bojanowski, Christian Puhrsch, and Armand Joulin. 2018. Advances in Pre-Training Distributed Word Representations. In Proceedings of the International Conference on Language Resources and Evaluation (LREC 2018).
    Google ScholarLocate open access versionFindings
  • Massimo Quadrana, Paolo Cremonesi, and Dietmar Jannach. 2018. SequenceAware Recommender Systems. ACM Comput. Surv. 51, 4, Article 66 (July 2018), 36 pages. https://doi.org/10.1145/3190616
    Locate open access versionFindings
  • Massimo Quadrana, Alexandros Karatzoglou, Balázs Hidasi, and Paolo Cremonesi. 2017. Personalizing session-based recommendations with hierarchical recurrent neural networks. In Proceedings of the Eleventh ACM Conference on Recommender Systems. ACM, 130–137.
    Google ScholarLocate open access versionFindings
  • Steffen Rendle, Christoph Freudenthaler, Zeno Gantner, and Lars Schmidt-Thieme. 2009. BPR: Bayesian personalized ranking from implicit feedback. In Proceedings of the twenty-fifth conference on uncertainty in artificial intelligence. AUAI Press, 452–461.
    Google ScholarLocate open access versionFindings
  • Steffen Rendle, Christoph Freudenthaler, and Lars Schmidt-Thieme. 2010. Factorizing personalized markov chains for next-basket recommendation. In Proceedings of the 19th international conference on World wide web. ACM, 811–820.
    Google ScholarLocate open access versionFindings
  • Chuan Shi, Zhiqiang Zhang, Ping Luo, Philip S Yu, Yading Yue, and Bin Wu. 2015. Semantic path based personalized recommendation on weighted heterogeneous information networks. In Proceedings of the 24th ACM International on Conference on Information and Knowledge Management. ACM, 453–462.
    Google ScholarLocate open access versionFindings
  • Yizhou Sun and Jiawei Han. 2013. Mining heterogeneous information networks: a structural analysis approach. Acm Sigkdd Explorations Newsletter 14, 2 (2013), 20–28.
    Google ScholarLocate open access versionFindings
  • Yizhou Sun, Jiawei Han, Xifeng Yan, Philip S Yu, and Tianyi Wu. 2011. Pathsim: Meta path-based top-k similarity search in heterogeneous information networks. Proceedings of the VLDB Endowment 4, 11 (2011), 992–1003.
    Google ScholarLocate open access versionFindings
  • Zhu Sun, Jie Yang, Jie Zhang, Alessandro Bozzon, Long-Kai Huang, and Chi Xu. 2018. Recurrent knowledge graph embedding for effective recommendation. In Proceedings of the 12th ACM Conference on Recommender Systems. ACM, 297–305.
    Google ScholarLocate open access versionFindings
  • Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. 2017. Attention is all you need. In Advances in Neural Information Processing Systems. 6000–6010.
    Google ScholarLocate open access versionFindings
  • Hongwei Wang, Fuzheng Zhang, Xing Xie, and Minyi Guo. 2018. Dkn: Deep knowledge-aware network for news recommendation. In Proceedings of the 2018 World Wide Web Conference on World Wide Web. International World Wide Web Conferences Steering Committee, 1835–1844.
    Google ScholarLocate open access versionFindings
  • Hongwei Wang, Fuzheng Zhang, Miao Zhao, Wenjie Li, Xing Xie, and Minyi Guo. 2019. Multi-Task Feature Learning for Knowledge Graph Enhanced Recommendation. arXiv preprint arXiv:1901.08907 (2019).
    Findings
  • Xiang Wang, Dingxian Wang, Canran Xu, Xiangnan He, Yixin Cao, and Tat-Seng Chua. 2018. Explainable Reasoning over Knowledge Graphs for Recommendation. arXiv preprint arXiv:1811.04540 (2018).
    Findings
  • Svante Wold, Kim Esbensen, and Paul Geladi. 1987. Principal component analysis. Chemometrics and intelligent laboratory systems 2, 1-3 (1987), 37–52.
    Google ScholarLocate open access versionFindings
  • Feng Yu, Qiang Liu, Shu Wu, Liang Wang, and Tieniu Tan. 2016. A dynamic recurrent model for next basket recommendation. In Proceedings of the 39th
    Google ScholarLocate open access versionFindings
  • Xiao Yu, Xiang Ren, Yizhou Sun, Quanquan Gu, Bradley Sturt, Urvashi Khandelwal, Brandon Norick, and Jiawei Han. 2014. Personalized entity recommendation: A heterogeneous information network approach. In Proceedings of the 7th ACM international conference on Web search and data mining. ACM, 283–292.
    Google ScholarLocate open access versionFindings
  • Fuzheng Zhang, Nicholas Jing Yuan, Defu Lian, Xing Xie, and Wei-Ying Ma. 2016. Collaborative knowledge base embedding for recommender systems. In Proceedings of the 22nd ACM SIGKDD international conference on knowledge discovery and data mining. ACM, 353–362.
    Google ScholarLocate open access versionFindings
  • Yongfeng Zhang and Xu Chen. 2018. Explainable recommendation: A survey and new perspectives. arXiv preprint arXiv:1804.11192 (2018).
    Findings
  • Yuyu Zhang, Hanjun Dai, Chang Xu, Jun Feng, Taifeng Wang, Jiang Bian, Bin Wang, and Tie-Yan Liu. 2014. Sequential Click Prediction for Sponsored Search with Recurrent Neural Networks.. In AAAI, Vol. 14. 1369–1375.
    Google ScholarLocate open access versionFindings
  • Wayne Xin Zhao, Gaole He, Kunlin Yang, Hongjian Dou, Jin Huang, Siqi Ouyang, and Ji-Rong Wen. 2019. KB4Rec: A Data Set for Linking Knowledge Bases with Recommender Systems. Data Intelligence 1, 2 (2019), 121–136.
    Google ScholarLocate open access versionFindings
  • Chang Zhou, Jinze Bai, Junshuai Song, Xiaofei Liu, Zhengchao Zhao, Xiusi Chen, and Jun Gao. 2018. Atrank: An attention-based user behavior modeling framework for recommendation. In Thirty-Second AAAI Conference on Artificial Intelligence.
    Google ScholarLocate open access versionFindings
Author
Xiaowen Huang
Xiaowen Huang
Quan Fang
Quan Fang
Shengsheng Qian
Shengsheng Qian
Yan Li
Yan Li
Your rating :
0

 

Tags
Comments
小科