Learning Latent Perception Graphs for Personalized Unknowns Recommendation

The fast-growing online-learning platforms, which are very convenient and contain rich course resources, have attracted many users to explore new knowledge online. However, the learning quality of online-learning is generally not as effective as offline classes. In offline studies in classrooms, teachers can interact with students and teach students in accordance with personal aptitude from students' feedback in classes. Without such real-time interaction, it is difficult for users to be aware of personal unknowns. In this paper, we consider an important issue to discover “user unknowns” from the question-giving process in online-learning platforms. A novel personalized learning framework, called PagBay, is devised to recommend user unknowns in the iterative round-by-round strategy, which contributes to applications such as a conversational bot. The flow enables users to progressively discover their weakness and to help them progress. However, discovering personal unknowns is quite challenging in online-learning platforms. Even though solving the problem with previous recommender algorithms provides solutions, they often lead to suboptimal results for unknowns recommendation as they simply rely on the user ratings and contextual features of questions. Generally, questions are associated with perceptions, and mining the relationships among users, questions, and perceptions potentially provide the clue to the better unknowns recommendation. Therefore, in this paper, we develop a novel recommender framework by borrowing strengths from perception-aware graph embedding for learning user unknowns. Our experimental studies on real data show that the proposed framework can effectively discover user unknowns in online learning services.

[1]  John Riedl,et al.  Item-based collaborative filtering recommendation algorithms , 2001, WWW '01.

[2]  H. Sebastian Seung,et al.  Algorithms for Non-negative Matrix Factorization , 2000, NIPS.

[3]  Peyman Kabiri,et al.  A Novel Non-Negative Matrix Factorization Method for Recommender Systems , 2015 .

[4]  Mingzhe Wang,et al.  LINE: Large-scale Information Network Embedding , 2015, WWW.

[5]  Jing Huang,et al.  Improving the Recommendation of Collaborative Filtering by Fusing Trust Network , 2012, 2012 Eighth International Conference on Computational Intelligence and Security.

[6]  Stephen J. Wright,et al.  Hogwild: A Lock-Free Approach to Parallelizing Stochastic Gradient Descent , 2011, NIPS.

[7]  Chuang Lin,et al.  Graph regularized non-negative matrix factorization with sparse coding , 2013, 2013 IEEE China Summit and International Conference on Signal and Information Processing.

[8]  Alexander J. Smola,et al.  Reducing the sampling complexity of topic models , 2014, KDD.

[9]  Julian J. McAuley,et al.  VBPR: Visual Bayesian Personalized Ranking from Implicit Feedback , 2015, AAAI.

[10]  Geoffrey J. Gordon,et al.  Relational learning via collective matrix factorization , 2008, KDD.

[11]  Lars Schmidt-Thieme,et al.  BPR: Bayesian Personalized Ranking from Implicit Feedback , 2009, UAI.

[12]  Yifan Hu,et al.  Collaborative Filtering for Implicit Feedback Datasets , 2008, 2008 Eighth IEEE International Conference on Data Mining.

[13]  D. Manjunath,et al.  Optimal recommendation to users that react: Online learning for a class of POMDPs , 2016, 2016 IEEE 55th Conference on Decision and Control (CDC).

[14]  Douglas B. Terry,et al.  Using collaborative filtering to weave an information tapestry , 1992, CACM.

[15]  Jeffrey Dean,et al.  Distributed Representations of Words and Phrases and their Compositionality , 2013, NIPS.

[16]  Xiaojun Wu,et al.  Graph Regularized Nonnegative Matrix Factorization for Data Representation , 2017, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[17]  Jeffrey Dean,et al.  Efficient Estimation of Word Representations in Vector Space , 2013, ICLR.