Holistic Entropy Reduction for Collaborative Filtering

Abstract We propose a collaborative filtering (CF) method that uses behavioral data provided as propositions having the RDF-compliant form of (user X, likes, item Y ) triples. The method involves the application of a novel self-configuration technique for the generation of vector-space representations optimized from the information-theoretic perspective. The method, referred to as Holistic Probabilistic Modus Ponendo Ponens (HPMPP), enables reasoning about the likelihood of unknown facts. The proposed vector-space graph representation model is based on the probabilistic apparatus of quantum Information Retrieval and on the compatibility of all operators representing subjects, predicates, objects and facts. The dual graph-vector representation of the available propositional data enables the entropy-reducing transformation and supports the compositionality of mutually compatible representations. As shown in the experiments presented in the paper, the compositionality of the vector-space representations allows an HPMPP-based recommendation system to identify which of the unknown facts having the triple form (user X, likes, item Y ) are the most likely to be true in a way that is both effective and, in contrast to methods proposed so far, fully automatic.

[1]  John Riedl,et al.  Application of Dimensionality Reduction in Recommender System - A Case Study , 2000 .

[2]  Pawel Misiorek,et al.  Long-Tail Recommendation Based on Reflective Indexing , 2011, Australasian Conference on Artificial Intelligence.

[3]  Kristian Kersting,et al.  Multi-Relational Learning with Gaussian Processes , 2009, IJCAI.

[4]  Takeo Kanade,et al.  Maximum Entropy for Collaborative Filtering , 2004, UAI.

[5]  Bracha Shapira,et al.  Recommender Systems Handbook , 2015, Springer US.

[6]  Roberto Turrin,et al.  Performance of recommender algorithms on top-n recommendation tasks , 2010, RecSys '10.

[7]  I. Pitowsky.  Quantum Probability ― Quantum Logic , 1989 .

[8]  Trevor Cohen,et al.  Reflective Random Indexing and indirect inference: A scalable method for discovery of implicit connections , 2010, J. Biomed. Informatics.

[9]  John Woods,et al.  A Quantum Logic of Down Below , 2006 .

[10]  Thierry Paul,et al.  Quantum computation and quantum information , 2007, Mathematical Structures in Computer Science.

[11]  Geoffrey I. Webb,et al.  Encyclopedia of Machine Learning , 2011, Encyclopedia of Machine Learning.

[12]  Joshua B. Tenenbaum,et al.  Modelling Relational Data using Bayesian Clustered Tensor Factorization , 2009, NIPS.

[13]  Andrzej Szwabe,et al.  RSVD-based Dimensionality Reduction for Recommender Systems , 2011 .

[14]  Jonathan L. Herlocker,et al.  Evaluating collaborative filtering recommender systems , 2004, TOIS.

[15]  Pawel Misiorek,et al.  Collaborative Filtering Based on Bi-Relational Data Representation , 2013 .

[16]  Elizabeth L. Wilmer,et al.  Markov Chains and Mixing Times , 2008 .

[17]  Achim Rettinger,et al.  Modeling and Learning Context-Aware Recommendation Scenarios Using Tensor Decomposition , 2011, 2011 International Conference on Advances in Social Networks Analysis and Mining.

[18]  Edwin R. Hancock,et al.  Pattern Vectors from Algebraic Graph Theory , 2005, IEEE Trans. Pattern Anal. Mach. Intell..

[19]  Alexander Dekhtyar,et al.  Information Retrieval , 2018, Lecture Notes in Computer Science.

[20]  V. Climenhaga.  Markov chains and mixing times , 2013 .

[21]  Arjen P. de Vries,et al.  Relevance information: a loss of entropy but a gain for IDF? , 2005, SIGIR '05.

[22]  C. J. van Rijsbergen,et al.  The geometry of information retrieval , 2004 .

[23]  Christof Monz,et al.  Applying Maximum Entropy to Known-Item Email Retrieval , 2008, ECIR.

[24]  Hans-Peter Kriegel,et al.  A Three-Way Model for Collective Learning on Multi-Relational Data , 2011, ICML.

[25]  Gediminas Adomavicius,et al.  Toward the next generation of recommender systems: a survey of the state-of-the-art and possible extensions , 2005, IEEE Transactions on Knowledge and Data Engineering.