Published July 11, 2018 | Version v1
Conference paper Open

Multi-Modal Adversarial Autoencoders for Recommendations of Citations and Subject Labels

  • 1. Kiel University
  • 2. ZBW -- Leibniz Information Centre for Economics


We present multi-modal adversarial autoencoders for recommendation and evaluate them on two different tasks: citation recommendation and subject label recommendation.  We analyze the effects of adversarial regularization, sparsity, and different input modalities.  By conducting 408 experiments, we show that adversarial regularization consistently improves the performance of autoencoders for recommendation.  We demonstrate, however, that the two tasks differ in the semantics of item co-occurrence in the sense that item co-occurrence resembles relatedness in case of citations, yet implies diversity in case of subject labels.  Our results reveal that supplying the partial item set as input is only helpful, when item co-occurrence resembles relatedness.  When facing a new recommendation task it is therefore crucial to consider the semantics of item co-occurrence for the choice of an appropriate model.


© Lukas Galke | ACM 2018. This is the author's version of the work. It is posted here for your personal use. Not for redistribution. The definitive Version of Record was published in UMAP '18- Proceedings of the 26th Conference on User Modeling, Adaptation and Personalization,



Files (747.0 kB)

Name Size Download all
747.0 kB Preview Download

Additional details


MOVING – Training towards a society of data-savvy information professionals to enable open leadership innovation 693092
European Commission