Heterogeneous dyadic multi-task learning with implicit feedback

Simon Moura, Amir Asarbaev, Massih Reza Amini, Yury Maximov

    Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

    1 Citation (Scopus)


    In this paper we present a framework for learning models for Recommender Systems (RS) in the case where there are multiple implicit feedback associated to items. Based on a set of features, representing the dyads of users and items extracted from an implicit feedback collection, we propose a stochastic gradient descent algorithm that learn jointly classification, ranking and embeddings for users and items. Our experimental results on a subset of the collection used in the RecSys 2016 challenge for job recommendation show the effectiveness of our approach with respect to single task approaches and paves the way for future work in jointly learning models for multiple implicit feedback for RS.

    Original languageEnglish
    Title of host publicationNeural Information Processing - 25th International Conference, ICONIP 2018, Proceedings
    EditorsLong Cheng, Seiichi Ozawa, Andrew Chi Sing Leung
    PublisherSpringer Verlag
    Number of pages13
    ISBN (Print)9783030041816
    Publication statusPublished - 2018
    Event25th International Conference on Neural Information Processing, ICONIP 2018 - Siem Reap, Cambodia
    Duration: 13 Dec 201816 Dec 2018

    Publication series

    NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
    Volume11303 LNCS
    ISSN (Print)0302-9743
    ISSN (Electronic)1611-3349


    Conference25th International Conference on Neural Information Processing, ICONIP 2018
    CitySiem Reap


    • Dyadic prediction
    • Multiple implicit feedback
    • Muti-task learning
    • Recommendation systems


    Dive into the research topics of 'Heterogeneous dyadic multi-task learning with implicit feedback'. Together they form a unique fingerprint.

    Cite this