Search

Research and education in machine learning requires diverse, representative, and open datasets that contain sufficient samples to handle the necessary training, validation, and testing tasks. Currently, the Recommender Systems area includes a large number of subfields in which accuracy and beyond-accuracy quality measures are continuously being improved. To feed this research variety, it is both necessary and convenient to reinforce the existing datasets with synthetic ones. This paper proposes a Generative Adversarial Network (GAN)-based method to generate collaborative filtering datasets in a parameterized way by selecting their preferred number of users, items, samples, and stochastic variability. This parameterization cannot be performed using regular GANs. Our GAN model is fed with dense, short, and continuous embedding representations of items and users, instead of sparse, large, and discrete vectors, to ensure fast and accurate learning, as compared to the traditional approach based on large and sparse input vectors. The proposed architecture includes a DeepMF model to extract the dense user and item embeddings and a clustering process to convert the dense GAN generated samples to the discrete and sparse samples necessary to create each required synthetic dataset. The results from three different source datasets show adequate distributions and expected quality values and evolutions in the generated datasets compared to the source datasets. Synthetic datasets and source codes are available to researchers. [ABSTRACT FROM AUTHOR]

Existing works on session-based recommendation have shown the advantage in enhancing the prediction ability of recommendation with various deep learning techniques. However, the following challenges need to be addressed: (1) the hierarchy of item transition patterns is overlooked; (2) existing works fail to distinguish various factors of item transition within a single session for disentangling user intents. To cope with the above challenges, we propose a novel session based recommendation model called S elf-supervised D ual H ypergraph learning with I ntent D isentanglement model (SDHID). Specifically, we first propose a disentangled capsule hypergraph convolutional channel for ne-grained intent learning to capture the intra-session pattern. Accordingly, we introduce the hypergraph and capsule networks in disentangling to learn the item embedding for different factors, and then the representation of the intra-session pattern is obtained by aggregating item embedding with attention weights. Moreover, we build a novel dual–primal hypergraph convolutional channel by mapping the hypergraph to a dual–primal graph for learning the item transition pattern of inter-session. In addition, the above two channels are combined into a self-supervised contrastive learning framework by maximizing mutual information between the learned session representations. We unify the recommendation and the self-supervised tasks under a primary and auxiliary learning framework. The combined optimization of two tasks leads to a hierarchical joint learning item transition for intra- and inter-session. Extensive experiments on real datasets show that the proposed model outperforms several state-of-the-art models. [ABSTRACT FROM AUTHOR]

In light of the emergence of deep reinforcement learning (DRL) in recommender systems research and several fruitful results in recent years, this survey aims to provide a timely and comprehensive overview of recent trends of deep reinforcement learning in recommender systems. We start by motivating the application of DRL in recommender systems, followed by a taxonomy of current DRL-based recommender systems and a summary of existing methods. We discuss emerging topics, open issues, and provide our perspective on advancing the domain. The survey serves as introductory material for readers from academia and industry to the topic and identifies notable opportunities for further research. [ABSTRACT FROM AUTHOR]

Recommender systems (RSs) have been employed for many real-world applications including search engines, social networks, and information retrieval systems as powerful intelligent techniques for creating personalized recommendations. RSs mainly utilize different data resources such as user-item ratings, tag data, and social relationships in their recommendation process to overcome the data sparsity issue. One of the main challenges is how to integrate the results obtained by these data resources. Moreover, most of the existing RSs focus on the recommendation accuracy, and do not consider the reliability of the recommendations. To address these challenges, in this study, a reliable recommendation method is developed, which employs deep neural networks and reinforcement learning. The proposed method firstly uses a stacked denoising autoencoder model to obtain the latent features from the rating, tag, and trust data. The similarity value between each pair of users is then calculated in three different aspects according to the extracted latent features. An ensemble approach is developed by integrating the three similarities using reinforcement learning to determine the final similarity value. The ensemble similarity values are used in the rating prediction process, and the reliability of the predicted ratings is evaluated using an effective measure. Finally, a recommendation strategy is developed based on the integration of the predicted ratings and their reliability values. We evaluated the effectiveness of the proposed method by conducting numerous experiments on three datasets where the results show that the proposed method is far superior to other recommendation models in making more accurate recommendations. [ABSTRACT FROM AUTHOR]

Recommender systems are popularly used to deal with an information overload issue. Existing systems mainly focus on user–item interactions and semantic information derived from metadata of users and items to improve recommendation accuracy. Item images provide useful information to infer users' individual preferences, especially for those domains where visual factors are influential such as fashion items. However, this type of information has been ignored by most previous work. To bridge this gap and meet the requirements of performance from the aspects of Accuracy, Scalability, and Explainability evaluation metrics, this paper proposes a scalable and explainable visually-aware recommender system framework called SEV-RS. This framework contains a visually-augmented heterogeneous information network, a scalable meta-path feature extraction method for multi-hop relations, and a shallow explainable meta-path based Collaborative Filtering recommendation approach. We compared SEV-RS with the state-of-the-art models such as the deep learning model using Graph Attention Network on two real-world datasets and one synthetic dataset. The results show that SEV-RS produced more accurate and more explainable recommendations. Also, SEV-RS has substantially less computational time than the compared deep learning models. [ABSTRACT FROM AUTHOR]

In the recommendation domain, user rating data has high sparsity and the number of interaction information from each user is very uneven, which brings great technical challenge to designing high-quality personalized recommendation schemes. Although autoencoder-based recommendation models have achieved some success in handling sparse data, they do not deal well with the uneven distribution of the amount of interaction information among different users, which further prevents the improvement of recommendation performance. In this paper, we find that the vanishing gradient problem occurs in the process of training model with the data from active users and theoretically analyze its root cause. Then, an improved autoencoder is proposed to alleviate this problem in recommendation domain. Furthermore, based on the improved autoencoder, two recommendation schemes are presented for rating prediction task and top- N ranking task, respectively. Finally, the numerical experimental results demonstrate that the proposed recommendation schemes can achieve about 5% and 3% improvements in the rating prediction and top- N ranking, respectively. Therefore, the improved model can well cope with the challenges brought by data sparsity and uneven distribution of rating data and achieve good recommendation performance. [ABSTRACT FROM AUTHOR]

Sequential recommendations have become a focus of attention across the deep learning community owing to their fitness to the actual application scenario. Although recently we have witnessed a surge of work on sequential recommender systems, they are still insufficient in exploring and exploiting item-attribute relations to enhance prediction accuracy. In this work, we propose a novel technological framework, MIA-SR, for sequential recommendation (SR) by modeling and predicting user preferences with multiple item attributes (MIA). When modeling the dynamic behavior of a user, not only the item sequence but also the attribute sequence is used to generate the fused representation of users. Further, we propose using a graph convolution network on the item-attribute bipartite graph to enhance the representations of items and attribute entities. Moreover, MIA-SR is naturally empowered with a multi-tasking strategy to exploit inductive bias among different preference signals and enhance item recommendation. Extensive experiments on public benchmark datasets have verified the merits of MIA-SR. The source code and data are available at: https://github.com/619496775/MIA-SR. [ABSTRACT FROM AUTHOR]

Recommendation systems or recommender systems (RSs) are very popular in entertainment websites. With the combination of neural networks and collaborative filtering, Neural Collaborative Filtering (NCF) recommendation methods have shown their outperformance in making item suggestions. However, the lack of semantic relationships between objects makes the NCF unable to capture the complex user–item interactions. Moreover, traditional NCF is unable to capture the dynamic user preference over time. To address these issues, in this paper, we propose novel semantic-enhanced NCF models which are applied to movie rating prediction and movie recommendation. Therefore, MovieLens and IMDB datasets are taken into account as case studies. The proposed models are the integration of ontology-like modeling and deep learning for recommendation tasks into two parts:(1) building the semantic knowledge base for movies and (2) building the user behavior analytic model that has semantic knowledge inference on the knowledge base combined with the sequential preference learned from user sessions, input into the NCF module for making predictions or recommendations. Several experiments have been conducted to show their better recommendation performance than the traditional NCF model. [Display omitted] • Building the semantic knowledge base for enhancing deep learning models. • Building the user behavior analytic model with the semantic knowledge base. • Experiments show the proposed models are significantly better than the original NCF. [ABSTRACT FROM AUTHOR]

Matrix Factorization is a successful approach for generating an effective recommender system. However, most existing matrix factorization methods suffer from the sparsity and cold-start issues of the user-item matrix as the primary information source of recommender systems. Besides, they are not much scalable to apply to large real-world applications. A main idea to overcome the cold-start and sparsity issues is to use additional information sources such as user/item profiles or user reviews on items. In this paper, a novel Attention-based Deep Learning Recommender System, so-called ADLRS, is proposed to employ the information sources in the matrix factorization method framework. The proposed method uses a language model to represent contextual information such that important features are effectively embedded. Moreover, a deep autoencoder reduces the dimensionality of item vectors embedded by the language model. Then, these vectors are used as regularization terms in the matrix factorization framework to form an objective function. Then, an iterative algorithm is designed to solve the objective function and provide a method prediction of unknown rating values. Experimental results show that the proposed method achieves superior performance compared to other state-of-the-art ones in most cases. Moreover, the improvement rate for sparse datasets and cold items proves that the proposed method effectively deals with sparsity, cold start and scalability problems. • An attention-based deep learning recommender system called ADLRS is proposed. • ADLRS incorporates profiles of items with the framework of matrix factorization. • The BERT language model is used to represents item profiles in the form of vectors. • A deep autoencoder is used to extract effective features and reduce the dimensionality of vectors. • An iterative method is used to solve the objective function and predict unknown ratings. [ABSTRACT FROM AUTHOR]

Data sparsity and cold start are two critical issues which need to be addressed in recommender systems (RSs). Currently, most methods address these issues by applying user history files or some side information to improve the user model and complete the rating matrix. However, such methods cannot perform well when labeled data is scarce or unavailable. In this paper, we propose a dual learning-based recommendation approach (DLRA). DLRA can trigger initial recommendation and improve the quality of recommendations by using the duality characteristics of RSs, even when the available labeled information is scarce. Specifically, DLRA regards the recommendation task as two independent subtasks — primal task and dual task, and these two tasks show strong duality in DLRA. The primal task is item-centered which aims to find users who can rate high for items, while the dual task is user-centered that aims to recommend the most favorite items to users. These two tasks have strong dualities in terms of the recommendation space, selection probability and recommendation basis. Based on these dualities, we design three dual learning strategies to couple the whole recommendation process and realize the self-tuning and self-improvement of each task model, and finally optimize the whole recommendation model. Based on the dataset of Movielens and BookCrossing, we simulate data sparsity and cold start recommendation scenarios, the experimental results show that DLRA achieves substantial improvement when the labeled data is scare, and it outperforms other hybrid recommendation approaches and deep learning strategies with a smaller predictive error as well as better recommendation accuracy. [ABSTRACT FROM AUTHOR]

A significant remaining challenge for existing recommender systems is that users may not trust recommender systems for either inaccurate recommendation or lack of explanation. Thus, it becomes critical to embrace a trustworthy recommender system. This survey provides a systematic summary of three categories of trust issues in recommender systems: social-aware recommender systems, which leverage users' social trust relationships; robust recommender systems, which filter untruthful information, noises and enhance attack resistance; and explainable recommender systems, which provide explanations of the recommended items. We focus on the work based on deep learning techniques, which is an emerging area in the recommendation research. • A systemic survey of current deep learning-based trust-aware recommendation methods from three aspects: social-awareness, robustness, and explainability. • A comprehensive overview and comment of the key research issues, and state-of-the-art techniques. • A summary and discussion of promising future directions to addressing remaining challenges. [ABSTRACT FROM AUTHOR]

Most existing recommendation methods assume that all the items are provided by separate producers rather than users. However, it could be inappropriate in some recommendation tasks since users may generate some items. Considering the user–item generation relation may benefit recommender systems that only use implicit user–item interactions. However, it may suffer from a dramatic imbalance. The number of user–item generation relations may be far smaller than the number of user–item interactions because each item is generated by at most one user. At the same time, this item can be interacted with by many users. To overcome the challenging imbalance issue, we propose a novel Deep Interaction-Attribute-Generation (DIAG) model. It integrates the user–item interaction relation, the user–item generation relation, and the item attribute information into one deep learning framework. The novelty lies in the design of a new item–item co-generation network for modeling the user–item generation information. Then, graph attention network is adopted to learn the item feature vectors from the user–item generations and the item attribute information by considering the adaptive impact of one item on its co-generated items. Extensive experiments conducted on two real-world datasets confirm the superiority of the DIAG method. [ABSTRACT FROM AUTHOR]

Recently, a variety of review-based recommendation systems that incorporate the valuable information extracted from user-generated textual reviews into user and item modeling have been proposed. However, the existing recommendations normally model reviews at the sentence level. They ignore the modeling of aspect words in reviews, which fails to capture user preferences and item attributes in a fine-grained way. In addition, few studies consider constructing user–item interaction based on review information extracted from the aspect level. In this paper, we are motivated to propose an Aspect-Aware Dual Graph Convolutional Network (DualGCN). Specifically, we first design an Aspect-GCN layer to model the message diffusion of an aspect graph constructed from reviews, capturing the overall description of an aspect in all reviews. We then propose a UI-GCN layer to model a user's fine-grained preferences toward interacted items at the aspect level. Finally, we adopt a Factorization Machine model to accomplish the recommendation task. The experimental results demonstrate that our model significantly outperforms the related approaches w.r.t. the accuracy of both rating prediction and top-K ranking on Amazon and Yelp datasets. [ABSTRACT FROM AUTHOR]

Precise prescription of medication dosing is crucial to patients, especially among Intensive Care Unit (ICU) patients. However, improper administration of some sensitive therapeutic medications (e.g., heparin) might place patients at unneeded risk, even life-threatening. Numerous factors such as a patient's clinical phenotype, genotype, and environmental factors will affect the heparin dose response. As a result, it is challenging to prescribe the optimal initial dose of heparin. In this paper, an individualized dosing policy is proposed to determine the optimal initial dose and minimize the risk of mis-dosing, as well as preventing the patients from late complications associated with medications dosing. A latent batch-constrained deep reinforcement learning (RL) algorithm is proposed to guarantee the safety of the medication recommendation system. The agent can observe a latent representation of patents and generate medication dosing solutions in successive and limited action spaces. The individualized dosing policy aims to reduce the extrapolation errors in the off-policy algorithms, by evaluating over-dosing and under-dosing of heparin in patients. Our results evaluated on Medical Information Mart for Intensive Care III (MIMIC-III) database demonstrate that the latent batch-constrained RL algorithm can work effectively from the retrospective data, showing promise to be used in future medication dosing policies. [ABSTRACT FROM AUTHOR]

The behavior, mental-health, emotion, life choices, social nature, and thought patterns of an individual are revealed by personality. Cyber forensics, personalized services, recommender systems are some of the examples of automatic personality prediction. A deep learning based personality prediction system has been developed in this work. Facial and ambient features are extracted from the visual modality using Multi-task Cascaded Convolutional Networks (MTCNN) and ResNet, respectively; the audio features are extracted using the VGGish Convolutional Neural Networks (VGGish CNN), and the text features are extracted using n-gram Convolutional Neural Networks (CNN). The extracted features are then passed to a fully connected layer followed by sigmoid for the final output prediction. Finally, the text, visual and audio modalities are combined in different ways: (i) concatenation of features in multi-modal setting, and (ii) application of different attention mechanisms for fusing features. The dataset released in Chalearn-17 is used for evaluating the performance of the system. From the obtained results, it can be concluded that, the concatenation of features extracted from different modalities attains comparable results with the averaging method (late fusion). It is also shown that a hand full of images are enough for attaining comparable performance. [ABSTRACT FROM AUTHOR]

Travel recommendation is very critical to helping users quickly find products or services that they are interested in. The key to travel recommender systems is learning user shopping intentions, which are expressed through various supervision signals, such as the clicked products and their titles. Existing travel recommendation methods commonly infer user intentions from click behaviors on travel products. However, remarkable keywords in the product title, such as departure, destination, travel time, hotel, and transportation are paid less attention. To this end, we hypothesize that modeling click sequences and product keywords in title jointly would result in a more holistic representation of a product and towards more accurate recommendations. Thus, we propose a TRKG (short for T ravel R ecommendation with K eywords G eneration) model, which fulfills the travel recommendation and keywords generation tasks simultaneously. To generate explainable outputs, unlike most previous approaches that regard the product title as a hidden feature vector, TRKG regards keywords in the product title as an additional supervision signal. Meanwhile, TRKG integrates the long-term and short-term user preferences in the travel recommendation component and the keywords generation component. To evaluate the proposed model, we constructed datasets from a large tourism e-commerce website in China. Extensive experiments demonstrate that the proposed method yields significant improvements over state-of-the-art methods. [ABSTRACT FROM AUTHOR]

Knowledge graphs (KGs) can provide rich, structured information for recommendation systems as well as increase accuracy and perform explicit reasoning. Deep reinforcement learning (RL) has also sparked great interest in personalized recommendations. The combination of the two holds promise in carrying out interpretable causal inference procedures and improving the performance of graph-structured recommendation. However, most KG-based recommendation focus on rich semantic relationships between entities in a heterogeneous knowledge graph, and thus fail to fully make use of the image information corresponding to an entity. In order to address these issues, we proposed a novel Multi-modal Knowledge-aware Reinforcement Learning Network (MKRLN), which couples recommendation and interpretability by providing actual paths in multi-modal KG (MKG). The MKRLN can generate path representation by composing the structural and visual information of entities, and infers the underlying rational of agent-MKG interactions by leveraging the sequential dependencies within a path from the MKG. In addition, as KGs have too many attributes and entities, their combination with RL leads to too many action spaces and states in the reinforcement learning space, which complicates the search of action spaces. Furthermore, in order to solve this problem, we proposed a new hierarchical attention-path, which makes users focus their attention on the items they are interested in. This reduces the relations and entities in the KGs, which in turn reduces the action space and state in RL, shortens the path to the target entity, and improves the accuracy of recommendation. Our model has explicit explanation ability in knowledge and images. Finally, we extensively evaluated our model on several large-scale real-world benchmark datasets, and it yielded favorable results compared with state-of-the-art methods. [ABSTRACT FROM AUTHOR]