logo

SCIENTIA SINICA Informationis, Volume 50 , Issue 12 : 1867(2020) https://doi.org/10.1360/SSI-2019-0177

Recurrent memory networks: modeling long short-term user preferences for session-based recommendation

More info
  • ReceivedAug 22, 2019
  • AcceptedJan 7, 2020
  • PublishedNov 20, 2020

Abstract


Funded by

国家自然科学基金(61832006,61872240)


References

[1] Zhou G, Zhu X, Song C, et al. Deep interest network for click-through rate prediction. In: Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, London, 2018. 1059--1068. Google Scholar

[2] Van den Oord A, Dieleman S, Schrauwen B. Deep content-based music recommendation. In: Advances in Neural Information Processing Systems, Lake Tahoe, 2013. 2643--2651. Google Scholar

[3] Wang H, Zhang F, Xie X, et al. Dkn: deep knowledge-aware network for news recommendation. In: Proceedings of the 2018 World Wide Web Conference, Lyon, 2018. 1835--1844. Google Scholar

[4] Lian J, Zhang F, Hou M, et al. Practical lessons for job recommendations in the cold-start scenario. In: Proceedings of the Recommender Systems Challenge 2017, Como, 2017. 1--6. Google Scholar

[5] Yang X, Liang C, Zhao M. Collaborative Filtering-Based Recommendation of Online Social Voting. IEEE Trans Comput Soc Syst, 2017, 4: 1-13 CrossRef Google Scholar

[6] Wang H, Wang J, Zhao M, et al. Joint topic-semantic-aware social recommendation for online voting. In: Proceedings of the 2017 ACM Conference on Information and Knowledge Management, Singapore, 2017. 347--356. Google Scholar

[7] Hidasi B, Karatzoglou A, Baltrunas L, et al. Session-based recommendations with recurrent neural networks. In: Proceedings of the 4th International Conference on Learning Representations, San Juan, 2016. Google Scholar

[8] Shani G, Heckerman D, Brafman R I. An MDP-based recommender system. J Mach Learn Res, 2005, 6: 1265--1295. Google Scholar

[9] Park S E, Lee S, Lee S g. Session-based collaborative filtering for predicting the next song. In: 2011 First ACIS/JNU International Conference on Computers, Networks, Systems and Industrial Engineering, Jeju, 2011. 353--358. Google Scholar

[10] Hidasi B, Quadrana M, Karatzoglou A, et al. Parallel recurrent neural network architectures for feature-rich session-based recommendations. In: Proceedings of the 10th ACM Conference on Recommender Systems, Boston, 2016. 241--248. Google Scholar

[11] Tan Y K, Xu X, Liu Y. Improved recurrent neural networks for session-based recommendations. In: Proceedings of the 1st Workshop on Deep Learning for Recommender Systems, Boston, 2016. 17--22. Google Scholar

[12] Greenstein-Messica A, Rokach L, Friedman M. Session-based recommendations using item embedding. In: Proceedings of the 22nd International Conference on Intelligent User Interfaces, Limassol, 2017. 629--633. Google Scholar

[13] Quadrana M, Karatzoglou A, Hidasi B, et al. Personalizing session-based recommendations with hierarchical recurrent neural networks. In: Proceedings of the 11th ACM Conference on Recommender Systems, Como, 2017. 130--137. Google Scholar

[14] Li J, Ren P, Chen Z, et al. Neural attentive session-based recommendation. In: Proceedings of the 2017 ACM Conference on Information and Knowledge Management, Singapore, 2017. 1419--1428. Google Scholar

[15] Jannach D, Lerche L, Jugovac M. Adaptation and evaluation of recommendations for short-term shopping goals. In: Proceedings of the 9th ACM Conference on Recommender Systems, Vienna, 2015. 211--218. Google Scholar

[16] Liu Q, Wu S, Wang L. Deepstyle: learning user preferences for visual recommendation. In: Proceedings of the 40th International ACM SIGIR Conference on Research and Development in Information Retrieval, Tokyo, 2017. 841--844. Google Scholar

[17] Ruocco M, Skrede O S L, Langseth H. Inter-session modeling for session-based recommendation. In: Proceedings of the 2nd Workshop on Deep Learning for Recommender Systems, Como, 2017. 24--31. Google Scholar

[18] Chen X, Xu H, Zhang Y, et al. Sequential recommendation with user memory networks. In: Proceedings of the 11th ACM International Conference on Web Search and Data Mining, Los Angeles, 2018. 108--116. Google Scholar

[19] Weston J, Chopra S, Bordes A. Memory networks. In: Proceedings of the 3rd International Conference on Learning Representations, San Diego, 2015. Google Scholar

[20] Sukhbaatar S, Weston J, Fergus R, et al. End-to-end memory networks. In: Proceedings of Advances in Neural Information Processing Systems, Montreal, 2015. 2440--2448. Google Scholar

[21] Miller A, Fisch A, Dodge J, et al. Key-value memory networks for directly reading documents. 2016,. arXiv Google Scholar

[22] Huang J, Zhao W X, Dou H, et al. Improving sequential recommendation with knowledge-enhanced memory networks. In: Proceedings of the 41st International ACM SIGIR Conference on Research & Development in Information Retrieval, Ann Arbor, 2018. 505--514. Google Scholar

[23] Cho K, Van Merriënboer B, Gulcehre C, et al. Learning phrase representations using rnn encoder-decoder for statistical machine translation. 2014,. arXiv Google Scholar

[24] Rendle S, Freudenthaler C, Gantner Z, et al. BPR: Bayesian personalized ranking from implicit feedback. In: Proceedings of the 25th Conference on Uncertainty in Artificial Intelligence, Montreal, 2009. 452--461. Google Scholar

[25] Jannach D, Ludewig M. When recurrent neural networks meet the neighborhood for session-based recommendation. In: Proceedings of the 11th ACM Conference on Recommender Systems, Como, 2017. 306--310. Google Scholar

[26] Hamilton W, Ying Z, Leskovec J. Inductive representation learning on large graphs. In: Proceedings of Advances in Neural Information Processing Systems, Long Beach, 2017. 1024--1034. Google Scholar

[27] Kingma D P, Ba J. Adam: a method for stochastic optimization. In: Proceedings of the 3rd International Conference on Learning Representations, San Diego, 2015. Google Scholar

[28] Tang Z, Wang D, Zhang Z. Recurrent neural network training with dark knowledge transfer. In: Proceedings of 2016 IEEE International Conference on Acoustics, Speech and Signal Processing, Shanghai, 2016. 5900--5904. Google Scholar

[29] Wang Q, Yin H, Hu Z, et al. Neural memory streaming recommender networks with adversarial training. In: Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, London, 2018. 2467--2475. Google Scholar

[30] Guo L, Yin H, Wang Q, et al. Streaming session-based recommendation. In: Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, Anchorage, 2019. 1569--1577. Google Scholar

[31] Li Z, Zhang Y, Wei Y, et al. End-to-end adversarial memory network for cross-domain sentiment classification. In: Proceedings of the 26th International Joint Conference on Artificial Intelligence, Melbourne, 2017. 2237--2243. Google Scholar

[32] Wang H, Zhang F, Wang J, et al. Ripplenet: propagating user preferences on the knowledge graph for recommender systems. In: Proceedings of the 27th ACM International Conference on Information and Knowledge Management, Turin, 2018. 417--426. Google Scholar

[33] Wang H, Zhang F, Wang J. Exploring High-Order User Preference on the Knowledge Graph for Recommender Systems. ACM Trans Inf Syst, 2019, 37: 1-26 CrossRef Google Scholar

[34] Wang H, Wang J, Wang J, et al. Graphgan: graph representation learning with generative adversarial nets. In: Proceedings of the 32nd AAAI Conference on Artificial Intelligence, New Orleans, 2018. 2508--2515. Google Scholar

[35] Wang H, Zhang F, Hou M, et al. Shine: signed heterogeneous information network embedding for sentiment link prediction. In: Proceedings of the 11th ACM International Conference on Web Search and Data Mining, Los Angeles, 2018. 592--600. Google Scholar

[36] Kipf T N, Welling M. Semi-supervised classification with graph convolutional networks. In: Proceedings of the 5th International Conference on Learning Representations, Toulon, 2017. Google Scholar

[37] Ying R, He R, Chen K, et al. Graph convolutional neural networks for web-scale recommender systems. In: Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, London, 2018. 974--983. Google Scholar

[38] Wang H, Zhao M, Xie X, et al. Knowledge graph convolutional networks for recommender systems. In: Proceedings of the World Wide Web Conference, San Francisco, 2019. 3307--3313. Google Scholar

[39] Wang H, Zhang F, Zhang M, et al. Knowledge-aware graph neural networks with label smoothness regularization for recommender systems. In: Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, Anchorage, 2019. 968--977. Google Scholar

  • Figure 1

    (Color online) The overall framework of RMN for two consecutive sessions

  • Figure 2

    (Color online) Illustration of (a) intra-session memory reader and (b) inter-session memory writer in RMN

  • Figure 3

    (Color online) Results of (a) Precision@$K$, (b) Recall@$K$, and (c) F1@$K$ on XING dataset. Some of the baselines are not shown in this figure for the sake of clarity

  • Figure 4

    (Color online) The impact of the length of users' historical sessions used in prediction. (a) Recall@K; protectłinebreak (b) MRR@K

  • Figure 5

    (Color online) Parameters sensitivity of RMN. (a) The number of slots of preference memory; (b) the dimension of preference memory

  • Table 1   Results of MRR@$K$ on MovieLens-20M and XING dataset
    MovieLens-20MXING
    MRR@2 MRR@5 MRR@10 MRR@20 MRR@50 MRR@2 MRR@5 MRR@10 MRR@20 MRR@50
    RMN 0.2737 0.2892 0.2940 0.2981 0.3055 0.0922 0.1085 0.1158 0.1188 0.1211
    M-Pop 0.0045 0.0073 0.0096 0.0130 0.0142 0.0003 0.0006 0.0010 0.0011 0.0014
    S-Pop 0.1077 0.1251 0.1430 0.1612 0.1727 0.0693 0.0923 0.0953 0.0954 0.0960
    Item-KNN 0.1855 0.2076 0.2196 0.2231 0.2249 0.0244 0.0380 0.0427 0.0453 0.0469
    Session-KNN 0.2014 0.2175 0.2216 0.2360 0.2442 0.0457 0.0525 0.0688 0.0742 0.0863
    RUM 0.0754 0.0846 0.1021 0.1068 0.1124 0.0205 0.0368 0.0425 0.0436 0.0477
    NARM 0.2512 0.2687 0.2745 0.2895 0.2911 0.0657 0.0884 0.0951 0.0977 0.0995
    RNN 0.2241 0.2389 0.2412 0.2426 0.2467 0.0512 0.0770 0.0867 0.0891 0.0904
    II-RNN 0.2354 0.2546 0.2753 0.2778 0.2810 0.0815 0.0867 0.0916 0.0974 0.1058
    HRNN 0.2310 0.2487 0.2502 0.2593 0.2644 0.0603 0.0883 0.0960 0.0992 0.1009
    GNN 0.2134 0.2271 0.2429 0.2533 0.2631 0.0512 0.0721 0.0767 0.0836 0.0882
  • Table 2   The comparison of RMN with its short-term-only (removing session initializer and inter-session memory writer) and long-term-only (removing intra-session memory reader) variants on XING dataset evaluated by F1@$K$
    F1@2 F1@5 F1@10 F1@20 F1@50
    RMN 0.0743 0.0571 0.0415 0.0261 0.0136
    RMN-short 0.0345 ($-53.6%$) 0.0408 ($-28.5%$) 0.0291 ($-29.9%$) 0.0190 ($-27.2%$) 0.0088 ($-35.3%$)
    RMN-long 0.0411 ($-44.7%$) 0.0455 ($-20.3%$) 0.0331 ($-20.2%$) 0.0220 ($-15.7%$) 0.0114 ($-16.2%$)