AI帮你理解科学

AI 生成解读视频

AI抽取解析论文重点内容自动生成视频


pub
生成解读视频

AI 溯源

AI解析本论文相关学术脉络


Master Reading Tree
生成 溯源树

AI 精读

AI抽取本论文的概要总结


微博一下
We introduced a scalable translation-based method, TransRec, for modeling the semantically complex personalized sequential dynamics in recommender systems

Translation-based Recommendation: A Scalable Method for Modeling Sequential Behavior.

IJCAI, pp.5264-5268, (2018)

被引用17|浏览96
EI
下载 PDF 全文
引用
微博一下

摘要

Modeling the complex interactions between users and items is at the core of designing successful recommender systems. One key task consists of predicting users’ personalized sequential behavior, where the challenge mainly lies in modeling ‘third-order’ interactions between a user, her previously visited item(s), and the next item to consu...更多

代码

数据

0
简介
  • In order to predict sequential user actions like the product to purchase, movie to watch, or place to visit, it is essential to model the third-order interactions between a user u, the item(s) i she recently consumed, and the item j to visit next.
  • Traditional recommendation methods usually excel at modeling two-way interactions
  • This includes Matrix Factorization (MF) techniques [Koren et al, 2009] that make use of inner products to model the compatibility between user-item pairs.
  • FPMC models third-order relationships between u, i, and j by the summation of two pairwise relationships: one for the compatibility between u and the item j, and another for the sequential continuity between the previous item i and the item j.
重点内容
  • In order to predict sequential user actions like the product to purchase, movie to watch, or place to visit, it is essential to model the third-order interactions between a user u, the item(s) i she recently consumed, and the item j to visit
  • Traditional recommendation methods usually excel at modeling two-way interactions. This includes Matrix Factorization (MF) techniques [Koren et al, 2009] that make use of inner products to model the compatibility between user-item pairs
  • Our work differs from the above in that we introduce a translation-based structure which naturally models the thirdorder interactions between a user, the previous item, and the item for personalized Markov transitions
  • We report the performance of each method on the test set in terms of the following ranking metrics: Area Under the ROC Curve (AUC): AUC = |U |
  • We introduced a scalable translation-based method, TransRec, for modeling the semantically complex personalized sequential dynamics in recommender systems
方法
  • Bayesian Personalized Ranking (BPR-MF) [Rendle et al, 2009]: BPR-MF is a state-of-the-art item recommendation model which takes Matrix Factorization as the underlying predictor.
  • It ignores sequential signals in the system.
  • Factorized Personalized Markov Chain (FPMC) [Rendle et al, 2010]: Uses a predictor that combines Matrix Factorization and factorized Markov Chains so that personalized Markov behavior can be captured (see Eq (2)).
  • In experiments the authors try both L1 and squared L2 distance for the predictor (see Eq (1))
结果
  • Evaluation Methodology

    For each dataset, the authors partition the sequence Su for each user u into three parts: (1) the most recent one S|uSu| for test, (2) the second most recent one S|uSu|−1 for validation, and (3) all the rest for training.
  • 1(Ru,gu ≤ 50), u∈U where gu is the ‘ground-truth’ item associated with user u at the most recent time step, Ru,i is the rank of item i for user u, and 1(b) is an indicator function that returns 1 if the argument b is true; 0 otherwise.
  • The main findings are summarized as follows: BPR-MF and FMC achieve considerably better results than the popularity-based baseline in most cases.
  • HRM achieves strong results amongst all baselines in many cases, presumably from the aggregation operations
结论
  • The authors introduced a scalable translation-based method, TransRec, for modeling the semantically complex personalized sequential dynamics in recommender systems.
  • The authors analyzed the connections between TransRec and existing methods, and demonstrated its suitability for modeling third-order interactions between users, their previously consumed items, and their item.
  • Superior results achieved on a spectrum of large, real-world datasets suggest that translation-based architectures are a promising avenue for recommendation problems
总结
  • Introduction:

    In order to predict sequential user actions like the product to purchase, movie to watch, or place to visit, it is essential to model the third-order interactions between a user u, the item(s) i she recently consumed, and the item j to visit next.
  • Traditional recommendation methods usually excel at modeling two-way interactions
  • This includes Matrix Factorization (MF) techniques [Koren et al, 2009] that make use of inner products to model the compatibility between user-item pairs.
  • FPMC models third-order relationships between u, i, and j by the summation of two pairwise relationships: one for the compatibility between u and the item j, and another for the sequential continuity between the previous item i and the item j.
  • Objectives:

    The authors aim to tackle the above issues by introducing a new framework called Translation-based Recommendation (TransRec).
  • Given the sequence set from all users S = {Su1 , Su2 , · · · , Su|U| }, the objective is to predict the item to be ‘consumed’ by each user and generate recommendation lists.
  • 3.1 The Proposed Model The authors aim to build a model that (1) naturally captures personalized sequential behavior, and (2) scales to large, realworld datasets
  • Methods:

    Bayesian Personalized Ranking (BPR-MF) [Rendle et al, 2009]: BPR-MF is a state-of-the-art item recommendation model which takes Matrix Factorization as the underlying predictor.
  • It ignores sequential signals in the system.
  • Factorized Personalized Markov Chain (FPMC) [Rendle et al, 2010]: Uses a predictor that combines Matrix Factorization and factorized Markov Chains so that personalized Markov behavior can be captured (see Eq (2)).
  • In experiments the authors try both L1 and squared L2 distance for the predictor (see Eq (1))
  • Results:

    Evaluation Methodology

    For each dataset, the authors partition the sequence Su for each user u into three parts: (1) the most recent one S|uSu| for test, (2) the second most recent one S|uSu|−1 for validation, and (3) all the rest for training.
  • 1(Ru,gu ≤ 50), u∈U where gu is the ‘ground-truth’ item associated with user u at the most recent time step, Ru,i is the rank of item i for user u, and 1(b) is an indicator function that returns 1 if the argument b is true; 0 otherwise.
  • The main findings are summarized as follows: BPR-MF and FMC achieve considerably better results than the popularity-based baseline in most cases.
  • HRM achieves strong results amongst all baselines in many cases, presumably from the aggregation operations
  • Conclusion:

    The authors introduced a scalable translation-based method, TransRec, for modeling the semantically complex personalized sequential dynamics in recommender systems.
  • The authors analyzed the connections between TransRec and existing methods, and demonstrated its suitability for modeling third-order interactions between users, their previously consumed items, and their item.
  • Superior results achieved on a spectrum of large, real-world datasets suggest that translation-based architectures are a promising avenue for recommendation problems
表格
  • Table1: Ranking results on different datasets (higher is better). The number of latent dimensions K for all comparison methods is set to 10. The best performance in each case is underlined. The last column shows the percentage improvement of TransRec over the best baseline
Download tables as Excel
相关工作
  • General recommendation. Traditional approaches to recommendation ignore sequential signals in the system. Such systems focus on modeling user preferences, and typically rely on Collaborative Filtering (CF) techniques, especially Matrix Factorization (MF) [Ricci et al, 2011]. For implicit feedback data (like purchases, clicks, and thumbsup), point-wise (e.g. [Hu et al, 2008; Pan et al, 2008; Ning and Karypis, 2011]) and pairwise methods (e.g. [Rendle et al, 2009]) based on MF have been proposed. Sequential recommendation. Scalable sequential models usually rely on Markov Chains (MC) to capture sequential patterns (e.g. [Rendle et al, 2010; Wang et al, 2015; Feng et al, 2015]). Rendle et al proposed to factorize the third-order ‘cube’ that represents the transitions made by users among items. The resulting model, Factorized Personalized Markov Chains (FPMC), can be seen as a combination of MF and MC and achieves good performance for nextbasket recommendation.
引用论文
  • [Bordes et al., 2013] Antoine Bordes, Nicolas Usunier, Alberto Garcia-Duran, Jason Weston, and Oksana Yakhnenko. Translating embeddings for modeling multi-relational data. In NIPS, pages 2787–2795, 2013.
    Google ScholarLocate open access versionFindings
  • [Chen et al., 2012] Shuo Chen, Joshua L. Moore, Douglas Turnbull, and Thorsten Joachims. Playlist prediction via metric embedding. In KDD, pages 714–722, 2012.
    Google ScholarLocate open access versionFindings
  • [Feng et al., 2015] Shanshan Feng, Xutao Li, Yifeng Zeng, Gao Cong, Yeow Meng Chee, and Quan Yuan. Personalized ranking metric embedding for next new poi recommendation. In IJCAI, pages 2069–2075, 2015.
    Google ScholarLocate open access versionFindings
  • [Hidasi et al., 2016] Balazs Hidasi, Alexandros Karatzoglou, Linas Baltrunas, and Domonkos Tikk. Session-based recommendations with recurrent neural networks. In ICLR, 2016.
    Google ScholarLocate open access versionFindings
  • [Hsieh et al., 2017] Cheng-Kang Hsieh, Longqi Yang, Yin Cui, Tsung-Yi Lin, Serge J. Belongie, and Deborah Estrin. Collaborative metric learning. In WWW, pages 193–201, 2017.
    Google ScholarLocate open access versionFindings
  • [Hu et al., 2008] Yifan Hu, Yehuda Koren, and Chris Volinsky. Collaborative filtering for implicit feedback datasets. In ICDM, pages 263–272, 2008.
    Google ScholarLocate open access versionFindings
  • [Jing and Smola, 2017] How Jing and Alexander J Smola. Neural survival recommender. In WSDM, pages 515–524, 2017.
    Google ScholarLocate open access versionFindings
  • [Koren et al., 2009] Yehuda Koren, Robert Bell, and Chris Volinsky. Matrix factorization techniques for recommender systems. Computer, pages 30–37, 2009.
    Google ScholarLocate open access versionFindings
  • [Levandoski et al., 2012] Justin J. Levandoski, Mohamed Sarwat, Ahmed Eldawy, and Mohamed F. Mokbel. LARS: A locationaware recommender system. In ICDE, pages 450–461, 2012.
    Google ScholarLocate open access versionFindings
  • [Lin et al., 2015] Yankai Lin, Zhiyuan Liu, Maosong Sun, Yang Liu, and Xuan Zhu. Learning entity and relation embeddings for knowledge graph completion. In AAAI, pages 2181–2187, 2015.
    Google ScholarLocate open access versionFindings
  • [McAuley et al., 2015] Julian J. McAuley, Christopher Targett, Qinfeng Shi, and Anton van den Hengel. Image-based recommendations on styles and substitutes. In SIGIR, pages 43–52, 2015.
    Google ScholarLocate open access versionFindings
您的评分 :
0

 

标签
评论
小科