越來越多線上服務被推出，而推薦系統成了從大量資訊中有效地擷取資訊及使用者喜好並提供個人化推薦的重要功能。在這些線上平台上，使用者透過各種操作與物品互動。使用者可以點擊物品、瀏覽物品、購買物品、評比物品等等。各種行為各自含有有關使用者喜好的資訊且行為之間有強烈的關聯。雖然這些多行為資訊存在於許多線上系統上，但大多數方法只考慮最主要的行為(如: 購買行為)。近期，有些研究轉向解決多行為推薦問題並在與許多單行為的方法比較中展現了效果的改進。然而，在推薦領域上鮮少研究考慮到使用者負行為。
在本次研究中，我們預期利用使用者多行為及考慮使用者負行為以改善推薦的品質，並提出Collaborative Filtering based on Multi-task Learning with Negative User Behavior(CFMTN) 的新穎方法。藉由多任務學習，每個行為都被當成單一任務並同步學習這些任務以獲得更好的模型。此外，在基於轉換法的預測層中擷取了行為之間的關聯。最後，CFMTN 預測使用者透過目標行為與物品互動的可能性。在三種類型資料集的實驗中，顯示CFMTN 結合使用者負面行為改善了推薦的品質。

More and more online services have been published, and recommendation system plays a key role to provide personalized recommendation by effectively extracting useful information and capturing user preference in the large volume of data. On these online platforms, users interact with items through various actions. Users can click an item, browse an item, buy an item, rating an item and so on. Each type of behavior contains rich information of user preference and there exists strong relationships among these behaviors. Although multiple types of user behaviors are ubiquitous in many online systems, most methods only consider the target behavior
(e.g., buy). Recently, some works turn to solve the multi-behavior recommendation
issue and show the performance improvements over several approaches on single behavior. However, little work on recommendation focuses on the negative user behavior.
In this study, we aim to improves the quality of recommendation with leveraging multiple user behaviors and considering the negative user behavior, and propose a novel method named Collaborative Filtering based on Multitask Learning with Negative User Behavior (CFMTN). With multitask learning, each type of behavior is treated as different task and all tasks are trained simultaneously to obtain a better model. In addition, the relations between different behaviors are captured in the transfer-based prediction layer. Eventually, CFMTN estimates the likelihood that user will interact with the item under the target behavior. Extensive experiments on real-world datasets demonstrate that CFMTN, which incorporates the negative user behavior, improves the quality of recommendation.

[1] X. He, L. Liao, H. Zhang, L. Nie, X. Hu, and T.­S. Chua, “Neural collaborative filtering,” in Proceedings of the 26th international conference on world wide web, pp. 173–182, 2017.
[2] X. Wang, X. He, M. Wang, F. Feng, and T.­S. Chua, “Neural graph collaborative filtering,” in Proceedings of the 42nd international ACM SIGIR conference on Research and development in Information Retrieval, pp. 165–174, 2019.
[3] S. Sedhain, A. K. Menon, S. Sanner, and L. Xie, “Autorec: Autoencoders meet collaborative filtering,” in Proceedings of the 24th international conference on World Wide Web, pp. 111–112, 2015.
[4] D. Liang, R. G. Krishnan, M. D. Hoffman, and T. Jebara, “Variational autoencoders for collaborative filtering,” in Proceedings of the 2018 world wide web conference, pp. 689–698, 2018.
[5] Q.­T. Truong, A. Salah, and H. W. Lauw, “Bilateral variational autoencoder for collaborative filtering,” in Proceedings of the 14th ACM International Conference on Web Search and Data Mining, pp. 292–300, 2021.
[6] C. Gao, X. He, D. Gan, X. Chen, F. Feng, Y. Li, T.­S. Chua, L. Yao, Y. Song, and D. Jin, “Learning to recommend with multiple cascading behaviors,” IEEE Transactions on Knowledge and Data Engineering, 2019.
[7] C. Chen, M. Zhang, W. Ma, Y. Zhang, Y. Liu, and S. Ma, “Efficient heterogeneous collaborative filtering without negative sampling for recommendation,” in Thirty­Fourth AAAI Conference on Artificial Intelligence, 2020.
[8] D. Song and D. A. Meyer, “Recommending positive links in signed social networks by optimizing a generalized auc,” in Twenty-­ninth AAAI conference on artificial intelligence, 2015.
[9] B. Paudel, T. Haas, and A. Bernstein, “Fewer flops at the top: Accuracy, diversity, and regularization in two­-class collaborative filtering,” in Proceedings of the Eleventh ACM Conference on Recommender Systems, pp. 215–223, 2017.
[10] B. Paudel, S. Luck, and A. Bernstein, “Loss aversion in recommender systems: Utilizing negative user preference to improve recommendation quality,” CoRR, vol. abs/1812.11422, 2018.
[11] H. Wen, L. Yang, and D. Estrin, “Leveraging post­-click feedback for content recommendations,” in Proceedings of the 13th ACM Conference on Recommender Systems, pp. 278–286, 2019.
[12] Y. Koren, R. Bell, and C. Volinsky, “Matrix factorization techniques for recommender systems,” Computer, vol. 42, no. 8, pp. 30–37, 2009.
[13] J. Davidson, B. Liebald, J. Liu, P. Nandy, T. Van Vleet, U. Gargi, S. Gupta, Y. He, M. Lambert, B. Livingston, et al., “The youtube video recommendation system,” in Proceedings of the fourth ACM conference on Recommender systems, pp. 293–296, 2010.
[14] Z. Cheng, Y. Ding, L. Zhu, and M. Kankanhalli, “Aspect­aware latent factor model: Rating prediction with ratings and reviews,” in Proceedings of the 2018 world wide web conference, pp. 639–648, 2018.
[15] X. Wang, X. He, L. Nie, and T.­S. Chua, “Item silk road: Recommending items from information domains to social users,” in Proceedings of the 40th International ACM SIGIR conference on Research and Development in Information Retrieval, pp. 185–194, 2017.
[16] D. P. Kingma and M. Welling, “Auto­encoding variational bayes,” in International Conference on Learning Representations (ICLR), 2013.
[17] A. P. Singh and G. J. Gordon, “Relational learning via collective matrix factorization,” in Proceedings of the 14th ACM SIGKDD international conference on Knowledge discovery and data mining, pp. 650–658, 2008.
[18] Z. Zhao, Z. Cheng, L. Hong, and E. H. Chi, “Improving user topic interest profiles by behavior factorization,” in Proceedings of the 24th International Conference on World Wide Web, pp. 1406–1416, 2015.
[19] B. Loni, R. Pagano, M. Larson, and A. Hanjalic, “Bayesian personalized ranking with multi­channel user feedback,” in Proceedings of the 10th ACM Conference on Recommender Systems, pp. 361–364, 2016.
[20] S. Rendle, C. Freudenthaler, Z. Gantner, and L. Schmidt­Thieme, “Bpr: Bayesian personalized ranking from implicit feedback,” in Proceedings of the Twenty­-Fifth Conference on Uncertainty in Artificial Intelligence, UAI ’09, (Arlington, Virginia, USA), p. 452–461, AUAI Press, 2009.
[21] J. Ding, G. Yu, X. He, Y. Quan, Y. Li, T.­S. Chua, D. Jin, and J. Yu, “Improving implicit recommender systems with view data.,” in IJCAI, pp. 3343–3349, 2018.
[22] L. Xia, C. Huang, Y. Xu, P. Dai, B. Zhang, and L. Bo, “Multiplex behavioral relation learning for recommendation via memory augmented transformer network,” in Proceedings of the 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval, pp. 2397–2406, 2020.
[23] C.­K. Hsieh, L. Yang, Y. Cui, T.­Y. Lin, S. Belongie, and D. Estrin, “Collaborative metric learning,” in Proceedings of the 26th international conference on world wide web, pp. 193–201, 2017.
[24] W. Wang, F. Feng, X. He, L. Nie, and T.­S. Chua, “Denoising implicit feedback for recommendation,” in Proceedings of the 14th ACM International Conference on Web Search and Data Mining, pp. 373–381, 2021.
[25] X. Wang, H. Fang, and C. Zhai, “A study of methods for negative relevance feedback,” in Proceedings of the 31st annual international ACM SIGIR conference on Research and development in information retrieval, pp. 219–226, 2008.
[26] S. Rendle, W. Krichene, L. Zhang, and J. Anderson, “Neural collaborative filtering vs. matrix factorization revisited,” in Fourteenth ACM Conference on Recommender Systems, pp. 240–248, 2020.
[27] Y. Hu, Y. Koren, and C. Volinsky, “Collaborative filtering for implicit feedback datasets,” in 2008 Eighth IEEE International Conference on Data Mining, pp. 263–272, Ieee, 2008.
[28] L. Xiao, Z. Min, Z. Yongfeng, L. Yiqun, and M. Shaoping, “Learning and transferring social and item visibilities for personalized recommendation,” in Proceedings of the 2017 ACM on Conference on Information and Knowledge Management, pp. 337–346, 2017.
[29] C. Chen, M. Zhang, Y. Zhang, Y. Liu, and S. Ma, “Efficient neural matrix factorization without sampling for recommendation,” ACM Transactions on Information Systems (TOIS), vol. 38, no. 2, pp. 1–28, 2020.
[30] C. Chen, M. Zhang, W. Ma, Y. Liu, and S. Ma, “Efficient non-­sampling factorization machines for optimal context­-aware recommendation,” in Proceedings of The Web Conference 2020, pp. 2400–2410, 2020.
[31] D. Liang, L. Charlin, J. McInerney, and D. M. Blei, “Modeling user exposure in recommendation,” in Proceedings of the 25th international conference on World Wide Web, pp. 951–961, 2016.
[32] X. He, H. Zhang, M.­Y. Kan, and T.­S. Chua, “Fast matrix factorization for online recommendation with implicit feedback,” in Proceedings of the 39th International ACM SIGIR conference on Research and Development in Information Retrieval, pp. 549–558, 2016.
[33] T. Evgeniou and M. Pontil, “Regularized multi–task learning,” in Proceedings of the tenth ACM SIGKDD international conference on Knowledge discovery and data mining, pp. 109–117, 2004.
[34] A. Evgeniou and M. Pontil, “Multi­task feature learning,” Advances in neural information processing systems, vol. 19, p. 41, 2007.
[35] J. Ni, J. Li, and J. McAuley, “Justifying recommendations using distantly­-labeled reviews and fine-grained aspects,” in Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLPIJCNLP), pp. 188–197, 2019.
[36] B. Brost, R. Mehrotra, and T. Jehan, “The music streaming sessions dataset,” in The World Wide Web Conference, pp. 2594–2600, 2019.
[37] C. C. Johnson, “Logistic matrix factorization for implicit feedback data,” Advances in Neural Information Processing Systems, vol. 27, no. 78, pp. 1–9, 2014.
[38] T. N. Kipf and M. Welling, “Semi-­supervised classification with graph convolutional networks,” in International Conference on Learning Representations (ICLR), 2017.
[39] X. Zhao, L. Zhang, Z. Ding, L. Xia, J. Tang, and D. Yin, “Recommendations with negative feedback via pairwise deep reinforcement learning,” in Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, pp. 1040–1048, 2018.
[40] M. Chen, A. Beutel, P. Covington, S. Jain, F. Belletti, and E. H. Chi, “Top-­k off­-policy correction for a reinforce recommender system,” in Proceedings of the Twelfth ACM International Conference on Web Search and Data Mining, pp. 456–464, 2019.