Social recommender systems (SocialRS) simultaneously leverage the user-to-item interactions as well as the user-to-user social relations for the task of generating item recommendations to users. Additionally exploiting social relations is clearly effective in understanding users’ tastes due to the effects of homophily and social influence. For this reason, SocialRS has increasingly attracted attention. In particular, with the advance of graph neural networks (GNN), many GNN-based SocialRS methods have been developed recently. Therefore, we conduct a comprehensive and systematic review of the literature on GNN-based SocialRS. In this survey, we first identify 84 papers on GNN-based SocialRS after annotating 2,151 papers by following the PRISMA framework (preferred reporting items for systematic reviews and meta-analyses). Then, we comprehensively review them in terms of their inputs and architectures to propose a novel taxonomy: (1) input taxonomy includes five groups of input type notations and seven groups of input representation notations; (2) architecture taxonomy includes eight groups of GNN encoder notations, two groups of decoder notations, and 12 groups of loss function notations. We classify the GNN-based SocialRS methods into several categories as per the taxonomy and describe their details. Furthermore, we summarize benchmark datasets and metrics widely used to evaluate the GNN-based SocialRS methods. Finally, we conclude this survey by presenting some future research directions. GitHub repository with the curated list of papers are available at https://github.com/claws-lab/awesome-GNN-social-recsys

Recently, deep learning has become a preferred choice for performing tasks in diverse application domains such as computer vision, natural language processing, sensor data analytics for healthcare, and collaborative filtering for personalized item recommendation. In addition, the Generative Adversarial Networks (GAN) has become one of the most popular frameworks for training machine learning models. Motivated by the huge success of GAN and deep learning on a wide range of fields, this paper explores an effective way to exploit both techniques into the collaborative filtering task for accurate recommendation. We have noticed that IRGAN and GraphGAN are pioneering methods that successfully apply GAN to recommender systems. However, we point out an issue regarding the employment of standard matrix factorization (MF) as their basic model, which is linear and unable to capture the non-linear, subtle latent factors underlying user-item interactions. Our proposed recommendation framework, named Collaborative Adversarial Autoencoders (CAAE), significantly extends the conventional IRGAN and GraphGAN as summarized below: (1) we use Autoencoder, which is one of the most successful deep neural networks, as our generator, instead of using the MF model; (2) we employ Bayesian personalized ranking (BPR) as our discriminative model; (3) we incorporate another generator model into our framework that focuses on generating negative items, which are items that a given user may not be interested in. We empirically test our framework using three real-life datasets along with four evaluation metrics. Owing to those extensions, our proposed framework not only produces considerably higher recommendation accuracy than the conventional GAN-based recommenders (i.e., IRGAN and GraphGAN), but also outperforms other state-of-the-art top-N recommenders (i.e., BPR, PureSVD, and FISM).

We address the multimedia recommendation problem, which utilizes items' multimodal features, such as visual and textual modalities, in addition to interaction information. While a number of existing multimedia recommender systems have been developed for this problem, we point out that none of these methods individually capture the influence of each modality at the interaction level. More importantly, we experimentally observe that the learning procedures of existing works fail to preserve the intrinsic modality-specific properties of items. To address above limitations, we propose an accurate multimedia recommendation framework, named MARIO, based on modality-aware attention and modality-preserving decoders. MARIO predicts users' preferences by considering the individual influence of each modality on each interaction while obtaining item embeddings that preserve the intrinsic modality-specific properties. The experiments on four real-life datasets demonstrate that MARIO consistently and significantly outperforms seven competitors in terms of the recommendation accuracy: MARIO yields up to 14.61% higher accuracy, compared to the best competitor.