Search

In this technical survey, the latest advancements in the field of recommender systems are comprehensively summarized. The objective of this study is to provide an overview of the current state-of-the-art in the field and highlight the latest trends in the development of recommender systems. It starts with a comprehensive summary of the main taxonomy of recommender systems, including personalized and group recommender systems. In addition, the survey analyzes the robustness, data bias, and fairness issues in recommender systems, summarizing the evaluation metrics used to assess the performance of these systems. Finally, it provides insights into the latest trends in the development of recommender systems and highlights the new directions for future research in the field. [ABSTRACT FROM AUTHOR]

Recommender systems are essential tools that provide personalized user experiences across various domains such as e-commerce, entertainment, social media, education and content streaming. The integration of auxiliary information, including user demographics, item attributes, and contextual data has shown significant promise in enhancing the performance of recommender systems. This systematic review investigates the impact of incorporating auxiliary information into various types of recommender systems, examining recent advancements, methodologies, datasets, evaluation metrics, and to equally examine its significance on generative artificial intelligence. Similarly, five (5) reputable online databases were used to identify the relevant studies for answering our research questions. To obtain effective results of our findings, we focus more on the recent studies published between (2019 - June 2024) to ensure that of our findings up-to-date. After filtering the selected primary papers that solely focused on auxiliary information recommender systems a total of 37 papers were identified and analyzed. Our analysis shows the most utilized datasets, metrics, models, addressed issues and future works. Research limitations and future scope are also highlighted to assist researchers and practitioners for their future studies.

With the rise of social media, individuals face challenges in decision-making due to the abundance of options available. Recommender Systems (RSs) leverage Artificial Intelligence (AI) to provide users with personalized suggestions aligned with their preferences and interests. This study presents a systematic review of AI-based Recommender Systems, focusing on recent advancements and primary studies published between 2019 and 2024. While several review papers have addressed various aspects of RSs, the rapid evolution of AI techniques necessitates an updated review to capture the latest trends and innovations. We systematically gathered data from five major databases: IEEE, Springer, Science Direct, ACM, and Wiley. Through the PRISMA methodology, we selected 85 relevant studies. Our analysis addresses several key research questions: the types of datasets and data sources used, major application fields, prevalent machine learning and AI techniques, overall research productivity, and the limitations and future trends in AI-based RSs. Our findings indicate that advanced AI techniques, particularly those incorporating deep learning with multiple hidden layers and transformer models like BERT, significantly enhance the accuracy and effectiveness of Recommender Systems. Furthermore, we observed a trend towards integrating contextual and real-time data to improve recommendation relevance. Additionally, we discuss ethical considerations such as privacy, data security, bias, and transparency, emphasizing the need for responsible AI development to ensure fair and equitable recommendations. These insights can guide future research and development efforts in the field.

In an era where digital information is abundant, the role of recommender systems in navigating this vast landscape has become increasingly vital. This study proposes a novel deep learning-based approach integrating multi-context and multi-criteria data within a unified neural network framework. The model processes these dimensions concurrently, significantly improving the precision of personalized recommendations. Context-aware and multi-criteria recommender systems extend traditional two-dimensional user-item preference methods with context awareness and multiple criteria. In contrast to traditional methods, our approach intricately weaves together multi-context and multi-criteria data within its architecture. This concurrent processing enables sophisticated interactions between context and criteria, enhancing recommendation accuracy. While context-aware systems incorporate contextual information such as time and location when making recommendations, multi-criteria-based approaches offer a spectrum of evaluative criteria, enriching the user experience with more tailored and relevant suggestions. Although both approaches have advantages in producing more accurate and personalized referrals, context information and multi-criteria ratings have not been employed together for producing recommendations. Our research proposes a novel deep learning-based approach for the multi-context, multi-criteria recommender system to address this gap. In contrast to traditional approaches that process context-aware recommender systems and multi-criteria recommender systems separately, our deep learning model intricately weaves together multi-context and multi-criteria data within its architecture. This integration is not staged; both dimensions are concurrently processed through a unified neural network framework. The model facilitates a sophisticated interaction between context and criteria by embedding these elements into the core of the network’s multiple layers. This methodology enhances the system’s adaptability and significantly improves its precision in delivering personalized recommendations, leveraging the compounded effects of contextual and criteria-specific insights. The proposed model shows superior performance in predictive tasks, achieving the lowest Mean Absolute Error (MAE) and Root Mean Square Error (RMSE) on the TripAdvisor and ITMRec datasets compared to other state-of-the-art recommendation techniques. Context-aware multi-criteria ratings data demonstrate the robustness and accuracy of the model.

In the realm of graph-based recommendation algorithms, the concept of metapaths plays a pivotal role in capturing complex semantic relationships between various types of entities, thereby significantly enhancing the recommendation quality. However, existing metapath-based recommendation algorithms often struggle to fully utilize the rich information embedded in metapaths while maintaining conciseness. Certain methods employ intricate and time-intensive aggregation networks to extract information from metapaths, while other strategies, prioritizing efficiency, neglect to investigate metapath information thoroughly. To address the above issues, we propose a Simple and Efficient Metapath Aggregated Network (SEMAN) for Recommendation. We employ a metapath pre-aggregation strategy using simple average pooling and concatenation operators to ensure the overall efficiency of the model. Furthermore, to fully exploit metapath information, the complete pathway nodes, including intermediate nodes, and metapath structural information are separately represented as aggregated vectors. Finally, we have designed a metapath semantic fusion module based on a dual-layer attention mechanism to process the metapath pre-aggregated results from both user and item sides across different granular levels. The theoretical analysis of time complexity substantiates the high efficiency of our model. Extensive experiments on three real-world graph datasets show that our proposed model performs significantly better than the state-of-the-art baselines on both Recall and NDCG metrics.

In the medical domain, particularly in intensive care units (ICUs), the immense volume of patient data presents a significant challenge for clinicians, often resulting in the oversight of critical information or excessive time consumption in accessing it. Recommender systems have been introduced to facilitate targeted, data-driven decision-making and ease the burden on healthcare professionals. This paper introduces Smart Summary, a novel distributed medical recommender system aimed at streamlining the analysis of extensive patient data and improving diagnostic accuracy by focusing on essential information. Smart Summary leverages patients’ admission reports and past lab values to predict International Classification of Diseases (ICD) codes, extract disease names, and forecast future abnormalities in lab values. Using this information, Smart Summary builds a comprehensive patient profile that covers the patient’s case precisely. Additionally, it recommends the most relevant laboratory values for individual patients by analyzing their data through various modules, including lab values abnormality prediction, automatic ICD codes prediction, and disease-named entity recognition. Furthermore, Smart Summary enhances its performance by incorporating doctors’ feedback, utilizing this information to refine recommendations for patients with similar profiles within the same cluster. Experimental results demonstrate that Smart Summary effectively learns to recommend relevant lab values for patients, achieving a Precision@10 of 0.92 after training on doctors’ feedback. Moreover, Smart Summary employs an efficient distributed machine learning method based on a split learning mechanism to ensure patient data privacy. This mechanism not only guarantees data privacy and security but also reduces communication overhead by 72% and computation overhead by 45.2% compared to the original split learning mechanism. To our knowledge, Smart Summary is the only system that creates comprehensive patient profiles using multiple machine learning models and recommends relevant lab values while ensuring privacy, efficiency, and security across various data sources.

Knowledge enhanced recommendation algorithms focus on how to leverage auxiliary information from knowledge graphs to enhance recommendation performance. However, existing methods for knowledge enhanced recommendation often overlook the issues of the non-uniform distribution of task-relevant information: (1) Item nodes often have neighbors unevenly distributed across interaction graphs and knowledge graphs. This uneven distribution of neighbor nodes might lead to the information from certain sources being ignored during message passing, thereby reducing the quality of the learned node embeddings. (2) The implicit inclusion of noise within graph data exacerbates the aforementioned issues, further hindering the effective utilization of knowledge graph information. In this paper, we introduce a novel algorithm called hierarchically coupled view-crossing contrastive learning for knowledge enhanced recommendation to address the challenges mentioned above. Specifically, we controllably couple knowledge graph information into each layer of message passing, and then use a weighted sum of the embeddings learned hierarchically as the final node representation. In addition, we devised a view-crossing contrastive learning approach to construct two additional contrastive learning loss functions for joint training with the main task and more effectively mitigate the adverse impact of noise than the traditional contrastive learning paradigms. Extensive experiments on three real-world graph datasets show that our proposed model performs significantly better than the state-of-the-art baselines and the results of experiments involving adversarial samples indicate the robustness of our model.

Graph neural networks have become a popular technique for collaborative filtering. However, most related work is based on user-item bipartite graphs, which can generate a large amount of noise due to the broad and elusive interests of users. To address this problem, we propose a novel generalized insertion framework (CGCN) that directly captures cliques in the item-item co-occurrence graph and considers them as the basic units of the user’s higher-order semantics. The method inserts the structural information in these item-item co-occurrence graphs as an insertion module into the original user-item bipartite graph propagation process, thus providing additional useful information to learn better feature representations. By utilizing the strong proximity relationships between different items in these cliques, the method is able to discover the user’s potential higher-order semantics. We experimentally evaluate two improved variants of the framework on three commonly used public datasets, and the results show significant performance improvements. The method is able to better discover users’ latent true intentions and achieve better recommender system performance by introducing clique information in the item-item co-occurrence graph.

In the age of information overload, recommender systems have emerged as essential tools, assisting users in decision-making processes by offering personalized suggestions. However, their effectiveness is contingent on the availability of large amounts of user data, raising significant privacy and security concerns. This review article presents an extended analysis of recommender systems, elucidating their importance and the growing apprehensions regarding privacy and data security. Federated Learning (FL), a privacy-preserving machine learning approach, is introduced as a potential solution to these challenges. Consequently, the potential benefits and implications of integrating FL with recommender systems are explored and an overview of FL, its types, and key components, are provided. Further, the privacy-preserving techniques inherent to FL are discussed, demonstrating how they contribute to secure recommender systems. By illustrating case studies and significant research contributions, the article showcases the practical feasibility and benefits of combining FL with recommender systems. Despite the promising benefits, challenges, and limitations exist in the practical deployment of FL in recommender systems. This review outlines these hurdles, bringing to light the security considerations crucial in this context and offering a balanced perspective. In conclusion, the article signifies the potential of FL in transforming recommender systems, paving the path for future research directions in this intersection of technologies.

While recommender systems have become an integral component of the Web experience, their heavy reliance on user data raises privacy and security concerns. Substituting user data with synthetic data can address these concerns, but accurately replicating these real-world datasets has been a notoriously challenging problem. Recent advancements in generative AI have demonstrated the impressive capabilities of diffusion models in generating realistic data across various domains. In this work we introduce a Score-based Diffusion Recommendation Module (SDRM), which captures the intricate patterns of real-world datasets required for training highly accurate recommender systems. SDRM allows for the generation of synthetic data that can replace existing datasets to preserve user privacy, or augment existing datasets to address excessive data sparsity. Our method outperforms competing baselines such as generative adversarial networks, variational autoencoders, and recently proposed diffusion models in synthesizing various datasets to replace or augment the original data by an average improvement of 4.30% in Recall@ $k$ and 4.65% in NDCG@ $k$ .

Taking advantage ofcontextual information can potentially boost the performance of recommender systems. In the era of Big Data, such side information often has several dimensions. Thus, developing decision-making algorithms to cope with such a high-dimensional context in real time is essential. That is specifically challenging when the decision-maker has a variety of items to recommend. In addition, changes in items' popularity or users' preferences can hinder the performance of the deployed recommender system due to a lack of robustness to distribution shifts in the environment. In this paper, we build upon the linear contextual multi-armed bandit framework to address this problem. We develop a decision-making policy for a linear bandit problem with high-dimensional feature vectors, a large set of arms, and non-stationary reward-generating processes. Our Thompson sampling-based policy reduces the dimension of feature vectors using random projection and uses exponentially increasing weights to decrease the influence of past observations with time. Our proposed recommender system employs this policy to learn the users' item preferences online while minimizing runtime. We prove a regret bound that scales as a factor of the reduced dimension instead of the original one. To evaluate our proposed recommender system numerically, we apply it to three real-world datasets. The theoretical and numerical results demonstrate the effectiveness of our proposed algorithm in making a trade-off between computational complexity and regret performance compared to the state-of-the-art.

The abundance of information on online purchasing websites makes it challenging for customers to locate products that match their preferences. However, the cold-start problem arises when there isn’t enough previous data, making it harder to make accurate recommendations for new customers or products. The enormous number of possible customers and products in a recommendation system leads to sparse data, which makes it harder to generate relevant recommendations and causes the sparsity problem. In addition, existing e-commerce recommender systems have difficulty making accurate product recommendations because they disregard individual consumer characteristics. In order to overcome these limitations, a hybrid recommender system combining ontology and sequential pattern mining (SPM) techniques is proposed. The strategy entails constructing an ontology that encompasses customer and product-related knowledge in the e-commerce domain. This ontology is then utilized to calculate customer preference similarities and generate predictions for the intended customer. The SPM algorithm is applied to the results of collaborative filtering to generate personalized recommendations for the customer. Experiments have demonstrated that the hybrid recommender system outperforms existing methods and resolves the cold-start problem and data sparsity in e-commerce recommender systems effectively. Even with limited initial data, the system generates accurate and individualized recommendations based on ontological domain knowledge and the customer’s sequential purchase patterns. By integrating ontology and sequential pattern mining, this strategy enhances the precision and individualization of the e-commerce industry’s recommendation process.

Students are a gold asset for each country. Proper guidance/recommendation to the students regarding their education-related issues can ultimately result in uplifting the economy of a country. Different education models are followed in the world, out of which Outcome Based Education (OBE) is the one. OBE education model comprises three main components, which include Program Educational Objectives (PEOs), Program Learning Outcomes (PLOs), and Course Learning Outcomes (CLOs). CLOs are outcomes that a student achieves after studying a course. A single course may contain one or more CLOs. These CLOs are then mapped to PLOs and PLOs are then mapped to PEOs. Therefore, our objective in this work is to improve deficient/weak CLOs of students by suggesting online resources. Whereas, in the absence of this proposed system, students have to find out these resources by themselves or course teacher recommends relevant online resources. To achieve this objective, we created a dataset for the OBE education model, as to date no standard dataset exists on the OBE education model. From this dataset, we created a Student-to-CLO matrix and performed bi-clustering on this matrix to find groups of students having similar performance in different CLOs. So far, Bi-clustering has been used in the Bio-informatics field to determine similarity in gene expression data. Generated bi-clusters are sorted according to their homogeneity of contained values. These sorted bi-clusters are then mapped to a 2D grid to formulate a reinforcement-learning environment. The start state of the recommendation agent is determined using cosine similarity. If an agent visits a state, deficient CLOs of that state are recommended to the student. The agent can visit only those states that are nearby to its current state and accessible through its legal action space. An optimal sequence of actions to visit different states of a 2D grid, which can improve student’s performance, is determined using Q-learning. Online resources including research articles, YouTube videos, books, and online tutorials are suggested to the student to improve his deficient CLOs using a mobile app.

On many online platforms, more items are being added every day, requiring users to select efficiently the desired item from among these candidates. In similar environments, recommender systems have been introduced, allowing users to select from among recommended items. While the immediate advantages of recommender systems, such as presenting users with pertinent information efficiently and boosting page views, are well recognized, concerns have been raised that these systems might constrict users’ choices and amplify echo chambers. Despite these concerns, the detrimental impacts of recommender systems on user behavior have not been thoroughly investigated. To address the societal challenges stemming from the limitation in the breadth of information to which users are exposed, a novel recommender system offering a wider array of choices is essential. In this study, we examine the user experience of the Nikkei Electronic Version, a major news delivery service in Japan, from the perspective of diversity. Using the language of article titles, we evaluate the diversity of recommended results and click logs presented to users. In addition to an analysis of the actual service, we conduct user experiments with participants recruited from among users of the service to investigate the effect of recommendation algorithms specialized in increasing diversity, separate from the actual service. We propose a recommender system based on the concept of extracting articles as far apart as possible within the language space. Through static experiments and user experiments, we show that the proposed recommender system is effective in diversifying the recommendation list. Furthermore, we show that when candidate articles that are likely to be clicked can be properly extracted, the proposed recommender system is effective in diversifying the articles that users click. Our research is aimed at improving long-term satisfaction by recommending content that users can enjoy in the short term while simultaneously ensuring diversity.

Recommender systems utilizing explicit feedback have witnessed significant advancements and widespread applications over the past years. However, generating recommendations in few-shot scenarios remains a persistent challenge. Recently, large language models (LLMs) have emerged as a promising solution for addressing natural language processing (NLP) tasks, thereby offering novel insights into tackling the few-shot scenarios encountered by explicit feedback-based recommender systems. To bridge recommender systems and LLMs, we devise a prompting template that generates user and item representations based on explicit feedback. Subsequently, we integrate these LLM-processed representations into various recommendation models to evaluate their significance across diverse recommendation tasks. Our ablation experiments and case study analysis collectively demonstrate the effectiveness of LLMs in processing explicit feedback, highlighting that LLMs equipped with generative and logical reasoning capabilities can effectively serve as a component of recommender systems to enhance their performance in few-shot scenarios. Furthermore, the broad adaptability of LLMs augments the generalization potential of recommender models, despite certain inherent constraints. We anticipate that our study can inspire researchers to delve deeper into the multifaceted dimensions of LLMs’ involvement in recommender systems and contribute to the advancement of the explicit feedback-based recommender systems field.

The fashion industry accounts for a relevant portion of the environmental impact of EU consumption. Moreover, the expansion of fast fashion raises further concerns about the well-being of the people and animals involved in its production. Increasing the purchase of green clothes (i.e., sustainable garments that have been produced by brands conforming to good ethical standards) is thus key to the reduction of fashion’s environmental and social footprint. In this paper, we investigate digital nudges to promote the selection of green clothes in online catalogs. For this purpose, we propose a recommender system that combines the personalized suggestion of new and second-hand garments with the presentation of sustainability and ethical standards data to favor item comparison and support responsible selection decisions. This is different from standard recommender systems, which suggest homogeneous products, either new or second-hand. Moreover, it is challenged by the tendency to buy new products observed in the literature about clothing consumption. In a user study involving 251 participants, we found that enhancing clothes recommendations with 1) sentences that promote second-hand garments and 2) visual labels that summarize items’ sustainability and brands’ ethical standards, sensibly reduced this tendency. Moreover, it induced some people to take the sustainability of products into account in their selection decisions. However, participants’ interest in brands’ ethical standards seemed to be secondary, especially regarding respect for animals. This finding reveals a need to enhance people’s awareness and sensitivity on this topic. Even though more work is needed to increase green, and especially ethical fashion consumption, our findings suggest the adoption of nudges in clothes recommender systems to enhance user awareness about items, their sustainability, and their social impact.

Modeling contextual information is a vital part of developing effective recommender systems. Still, existing work on recommendation algorithms has generally put limited focus on the effective treatment of contextual information. Moreover, adding context to recommendation models is challenging since it increases the dimensionality and complexity of the model. Therefore, an efficient learning method is required to extract an association and inter-relationship between user/item features and contextual features for preference-driven modeling. The engineering of features through the exploration of adjacent correlations between the user/item and their context, and their further learning through a distance-based metric, is critical for effective personalization. Motivated by this, we introduce a context-aware recommendation strategy using a ‘contextual grid triplet network.’ This strategy uses a contextual grid topology to capture robust semantic representations of users, items, and contextual data. We present a learning methodology that merges a triplet network with a convolutional neural network. This fusion enables the exploration of associations both ‘within’ the contextual grid, such as between users or items, and ‘between’ different contextual grids, like between a user and items of input. Moreover, we present a variant of a hinge loss function using a triplet network for improved performance and fast convergence. In this work, we study how these aspects boost the quality of top-N recommendations. Furthermore, We show through extensive ablation-based experiments that the proposed method outperforms existing state-of-the-art techniques, demonstrating its robustness and feasibility.

Learning the complex relationships between items in a sequential recommendation system (SRS) and session-based recommendation system (SBRS) is critical for obtaining higher prediction scores. In recent studies, to capture item-item information, items have been represented as the nodes of graph neural networks (GNNs) and the relevance of items with self-/soft attention layers has been calculated. GNNs have been used because standalone attention-based methods focus only on the relative significance of items within a single session, neglecting high-order item-item relationships that change through sessions. The relational summarization task is a natural language processing task that extracts the relationship between two tokens from a related corpus; however, its adaptation to SRS and SBRS is unknown. To fill this lacuna, in this study, the relationships between items from related sessions are extracted using the transformer-based abstractive summarization model PEGASUS. To improve session embedding, the proposed model, named “gap-session transformer” utilizes gap-session masking to learn the relationships between items within different sessions. In addition, a group of sessions are divided into multiple corpus sets based on the theme of each corpus, and the autoregressive beam-search decoder is connected to a transformer decoder for the generation of the next session while auxiliary tasks are performed to enhance the recommendation task. Extensive experiments conducted on the MovieLens1M dataset and Yoochoose dataset verify that our model significantly outperforms the state-of-the-art (SOTA) methods, and the results demonstrate the efficacy of the relational summarization task in recommendation systems.

Recommender systems are an effective way to address the information overload problem. Due to the rapid development of deep learning, recommender systems based on deep learning can efficiently process massive training samples and integrate a variety of additional information, which can alleviate the inherent data sparsity and cold start problems of recommender systems. However, deep learning recommender systems need to collect massive amounts of user data for training models, which poses challenges for data security and user privacy protection. By using users’ local, not centrally collected, data, federated learning can collaboratively train deep learning recommendation models and improve the accuracy of recommendations without compromising user privacy. In this paper, we propose FedDeepFM, a deep recommender system model based on federated learning. First, FedDeepFM follows the standard paradigm of federated learning by uploading each client’s model parameters instead of the original data for model updates, according to the data processing inequality, which ensures that the server cannot collect users’ private information directly but at the same time is able to use the private data held by each client to jointly train a global model with generalization capabilities. Second, to prevent the server from learning users’ private information through indirect inference methods, a pseudo-interaction filling method is used to expand the client data to mask the real user data. The experimental results show that FedDeepFM, the deep recommender system model based on federated learning proposed in this paper, can provide high-quality recommendations while enhancing users’ privacy.

Planning an itinerary is a complex activity, which includes the choice of a few places to see, coupled with information on timing, transferring methods and related activities. Intelligent tools such as recommender systems have been used in order to support these activities. While decisions regarding the type of trip to undertake are strongly influenced by tourists’ personalities, currently only a few recommenders exploit information about this aspect. Our aim is to provide itinerary recommender system designers with some guidance on the integration of knowledge on personality traits and itinerary factors in a recommender. To do so, first we modeled the most important aspects of an itinerary, starting from the state-of-the-art literature on recommender systems for tourism. We identified thirteen factors, from the variety (in type and topic) of Points-of-Interest (POIs) to the expected duration of transfer times, grouped into three broader dimensions (POIs, time and choice modality). Then, we carried out a survey-based study on Generation Z (namely, the generation of people born between 1996/1997 and 2012) to investigate if the Big Five personality traits can affect the user’s decision-making process when planning an itinerary, and, in particular if they are related to user preferences for the itinerary factors in our model. Finally, we used our findings to define some guidelines for the design of advanced itinerary recommender systems.

Researchers typically leverage side information, such as social networks or the knowledge graph, to overcome the sparsity and cold start problem in collaborative filtering. To tackle the limitations of existing user interest modeling, we propose a knowledge-enhanced user multi-interest modeling for recommender systems (KEMIM). First, we utilize the user-item historical interaction as the knowledge graph’s head entity to create a user’s explicit interests and leverage the relationship path to expand the user’s potential interests through connections in the knowledge graph. Second, considering the diversity of a user’s interests, we adopt an attention mechanism to learn the user’s attention to each historical interaction and each potential interest. Third, we combine the user’s attribute features with interests to solve the cold start problem effectively. With the knowledge graph’s structural data, KEMIM could describe the features of users at a fine granularity and provide explainable recommendation results to users. In this study, we conduct an in-depth empirical evaluation across three open datasets for two different recommendation tasks: Click-Through rate (CTR) prediction and Top-K recommendation. The experimental findings demonstrate that KEMIM outperforms several state-of-the-art baselines.

In this era of information overload, to better provide personalized content services to users, recommendation systems have greatly improved the efficiency of information distribution. Graph Convolution Network(GCN), which is one of the representative works of graph structure aggregation processing, works by node convolution with the help of the Laplacian matrix of the graph and weighted combination of neighbor node information according to the outgoing and incoming degrees of neighbor nodes to obtain the representation of the current node. However, the mainstream GCN models nowadays do not take into account data augmentation of metadata and the fact that each node plays different roles with different importance and weights, thus making the recommendation performance limited. To better solve the above problems, we propose the IcaGCN model, which can perform data augmentation and calculate node weights in modules, and is a convenient plug-and-play method. Finally, extensive experimental results on four real-world datasets have shown the effectiveness and robustness of the proposed model. Especially on the Amazon-Book dataset, our IcaGCN has improved by 6.32%, 42.29%, and 12.38% in Recall@20, MRR@20, and NDCG@20, respectively, compared to other existing state-of-the-art models. We also provide source code and data to reproduce the experimental results available at https://github.com/PersonZ1223/IcaGCN.git

The emergence of recommender system is aimed at solving the problems brought by information explosion to human life and even the development of human society. As a traditional recommendation technique, collaborative filtering often encounters sparsity and cold start problems in many recommendation scenarios. Therefore, researchers have found that the introduction of side information can solve these problems to a certain extent and improve the performance of recommender systems. The knowledge graph is a heterogeneous graph that contains rich semantic relationships among items. The Multi-Perspective Learning based on Transformer Knowledge Graph Enhanced Recommendation (MPL-TransKR) proposed in this paper uses the knowledge graph as the side information for input and introduces the multi-head self-attention mechanism by reasonably combining the transformer idea. While learning the high-order neighborhood information of the items, the long-distance information between the items is captured, and the weight value is assigned to the user through the attention mechanism to strengthen the user representation to realize user-item multi-perspective learning to enhance the performance of the recommendation model. Through extensive experiments using public datasets, we demonstrated that MPL-TransKR performs well in book and music recommendations, surpassing state-of-the-art baselines on several metrics.

Collaborative filtering methods for recommender systems tend to represent users as a single static latent vector. However, user behavior and interests may dynamically change in the context of the recommended item being presented to the user. For example, in the case of movie recommendations, it is usually true that movies that the user watched more recently are more informative than movies that were watched a long time ago. However, it is possible that a particular movie from the past may become suddenly more relevant for prediction in the presence of a recommendation for its sequel movie. In response to this issue, we introduce the Attentive Item2Vec++ (AI2V++) model, a neural attentive collaborative filtering approach in which the user representation adapts dynamically in the presence of the recommended item. AI2V++ employs a novel context-target attention mechanism in order to learn and capture different characteristics of the user’s historical behavior with respect to a potential recommended item. Furthermore, analysis of the neural-attentive scores allows for improved interpretability and explainability of the model. We evaluate our proposed approach on five publicly available datasets and demonstrate its superior performance in comparison to state-of-the-art baselines across multiple accuracy metrics.

Session-based recommendation is a crucial task aiming to predict users’ interested items based only on anonymous user behaviors. Most recent solutions for session-based recommendation comprehensively consider the interactive information of all sessions but bring the problem of imbalanced positive and negative samples on model training. In this paper, we propose a novel approach, named Attention-enhanced Graph Neural Networks with Global Context for Session-based Recommendation (AGNN-GC), to learn and merge item transitions of all sessions in a cleverer way to enhance the recommendation effects. AGNN-GC first constructs global and local graphs based on all training sequences. Next, it uses graph convolutional networks with a session-aware attention mechanism to learn global-level item embedding in all sessions. Then it employs a graph attention networks module to learn local-level item embedding in the current sessions. After that, it fuses the learned two-level item embedding to enhance the feature presentations of items in the current session by a novel attention mechanism. Finally, applying the focal loss to balance positive and negative samples on model training accomplishes the prediction. Our experiments on three real-world datasets consistently show the superior performance of AGNN-GC over state-of-the-art methods.

The way that computing systems digest geographic space is fundamentally different from people’s understanding of space. In human discourse, a geographic space is referred to by a place name, and the reasoning about a place are based on its characteristics. This is in contrast with computing systems where geographical spaces are handled by the definition of coordinate systems. Hence, when recommending places, a recommendation method that leverages textual content, as a medium of communication among people, can be better understood. In this paper, we use elements of Natural Language Processing (NLP), such as Positive Point-wise Mutual Information (PPMI), Term Frequency - Inverse Document Frequency (TF-IDF), and Multi-Dimensional Scaling (MDS), to infer a conceptual space of the items of a place-based recommender system. By applying a Support Vector Machine (SVM) classifier on the resulting conceptual space, some meaningful directions are extracted. Shannon entropy is used as a measure to identify the directions that imply a valid geographic region. We apply the method on a dataset of advertisement descriptions of rental properties and a dataset of Persian Wikipedia articles. The results showed the proposed method is able to measure the similarity of items in the inferred conceptual space with 88% of accuracy. A comparison with BERT algorithm demonstrates the superiority of the proposed method over the baseline models.

Extensive user check-in data incorporating user preferences for location is collected through Internet of Things (IoT) devices, including cell phones and other sensing devices in location-based social network. It can help traveling enterprises intelligently predict users' interests and preferences, provide them with scientific tourism paths, and increase the enterprises income. Thus, successive point-of-interest (POI) recommendation has become a hot research topic in augmented Intelligence of Things (AIoT). Presently, various methods have been applied to successive POI recommendations. Among them, the recurrent neural network-based approaches are committed to mining the sequence relationship between POIs, but ignore the high-order relationship between users and POIs. The graph neural network-based methods can capture the high-order connectivity, but it does not take the dynamic timeliness of POIs into account. Therefore, we propose an Interaction-enhanced and Time-aware Graph Convolution Network (ITGCN) for successive POI recommendation. Specifically, we design an improved graph convolution network for learning the dynamic representation of users and POIs. We also designed a self-attention aggregator to embed high-order connectivity into the node representation selectively. The enterprise management systems can predict the preferences of users, which is helpful for future planning and development. Finally, experimental results prove that ITGCN brings better results compared to the existing methods. [ABSTRACT FROM AUTHOR]

Convolutional neural network (CNN)-based recommender systems are playing an increasingly significant role in the vigorous development of Industrial Internet of Things, and have made great contributions to analyzing and mining a large amount of data to provide various services for terminal users. However, as the lack of explainability in deep learning, users often have low trust in the system due to their incomprehension of recommendation results. In addition, recommender systems have been facing a serious sparsity problem, and relying only on sparse rating data to learn user preferences and similarities may face malicious recommendation attacks. The abovementioned problems have been hindering the further improvement of recommendation performance. Therefore, in order to effectively alleviate the sparsity problem and meanwhile enhance the trustworthiness, an auxiliary review-based personalized attentional CNN (ARPCNN) is proposed in this article. By applying the proposed personalized word-level attention mechanism and personalized review-level attention mechanism in parallel CNNs, critical words and informative reviews are given high attention weights. Moreover, a user auxiliary network is proposed, which regards the reviews written by kindred spirits who have a trust relationship with the user as auxiliary reviews, and effectively extracts the user’s auxiliary review features, thereby achieving more accurate user modeling to improve the recommendation performance. Extensive experiments are conducted on four real-world datasets, and the results show that the performance of the proposed model is better than that of baselines, which verifies the effectiveness of ARPCNN. [ABSTRACT FROM AUTHOR]

The low-rank matrix completion has gained rapidly increasing attention from researchers in recent years for its efficient recovery of the matrix in various fields. Numerous studies have exploited the popular neural networks to yield low-rank outputs under the framework of low-rank matrix factorization. However, due to the discontinuity and nonconvexity of rank function, it is difficult to directly optimize the rank function via back propagation. Although a large number of studies have attempted to find relaxations of the rank function, e.g., Schatten- $p$ norm, they still face the following issues when updating parameters via back propagation: 1) These methods or surrogate functions are still non-differentiable, bringing obstacles to deriving the gradients of trainable variables. 2) Most of these surrogate functions perform singular value decomposition upon the original matrix at each iteration, which is time-consuming and blocks the propagation of gradients. To address these problems, in this paper, we develop an efficient block-wise model dubbed differentiable low-rank learning (DLRL) framework that adopts back propagation to optimize the Multi-Schatten- $p$ norm Surrogate (MSS) function. Distinct from the original optimization of this surrogate function, the proposed framework avoids singular value decomposition to admit the gradient propagation and builds a block-wise learning scheme to minimize values of Schatten- $p$ norms. Accordingly, it speeds up the computation and makes all parameters in the proposed framework learnable according to a predefined loss function. Finally, we conduct substantial experiments in terms of image recovery and collaborative filtering. The experimental results verify the superiority of the proposed framework in terms of both runtimes and learning performance compared with other state-of-the-art low-rank optimization methods. Our codes are available at https://github.com/chenzl23/DLRL. [ABSTRACT FROM AUTHOR]

In recommendation tasks, we model user preferences by learning node representations (i.e., user and item embeddings) based on the observed user-item interaction data, which is a bipartite graph. Graph Neural Networks (GNNs) are widely used to refine the representations by exploring the topology of the graph: embeddings of neighbors are propagated to each node to reconstruct its embeddings. However, the propagation strategy in existing GNNs is empirical and defective: (1) a substantial proportion of links are missed in the sparse observed graph, which causes ineffective and biased propagation; and (2) the propagation weights are determined by a coarse pre-defined rule, which only takes the degree of nodes into consideration. In this paper, we propose a dense and data-driven propagation mechanism for GNNs. Considering the graph we use to propagate embeddings in recommendation tasks is extremely sparse, we complement it and use the predicted graph as the new propagation tool. We learn the propagation matrix from the data, and propose a Self-propagation Graph Neural Network (SGNN). Since it is very space- and time-consuming to maintain a large and dense propagation matrix, we factorize it for storing and updating. In this paper, we propose three methods to complete the sparse graph and construct the propagation matrix: (1) we complete the graph based on a recommendation model; (2) we measure the node distance based on spectral clustering; (3) we predict missing links of the graph based on predictive embeddings. In SGNN, the embeddings can be propagated to not only the observed neighbors, but also the potential yet unobserved neighbors, and the propagation weights are learned based on the connection strength. Comprehensive experiments on three real-world datasets demonstrate the effectiveness and efficiency of our proposed model: SGNN outperforms recent state-of-the-art GNNs significantly. Codes are available on https://github.com/Wenhui-Yu/LCFN. [ABSTRACT FROM AUTHOR]

One-class collaborative filtering (OCCF) problems are ubiquitous in real-world recommendation systems, such as news recommendation, but suffer from data sparsity and lack of negative items. To address the challenge, the state-of-the-art algorithm assigns uninteracted items with smaller weights of being negative and performs low-rank approximation over the user-item interaction matrix. However, the prior ratings are usually suggested to be zero but may not be well-defined. To avert the direct utilization of prior ratings for uninteracted items, we propose a novel ranking-based implicit regularizer by hypothesizing that users’ preference scores for uninteracted items should not deviate a lot from each other. The regularizer is then used in a ranking-based OCCF framework to penalize large differences of preference scores between uninteracted items. To efficiently optimize model parameters in this framework, we develop the scalable alternating least square algorithm and coordinate descent algorithm, whose time complexity is linearly proportional to the data size. Finally, we extensively evaluate the proposed algorithms on six public real-world datasets. The results show that the proposed regularizer significantly improves the recommendation quality of ranking-based OCCF algorithms, such as BPRMF and RankALS. Moreover, the ranking-based framework with the proposed regularizer outperforms the state-of-the-art recommendation algorithms for implicit feedback. [ABSTRACT FROM AUTHOR]

As sharing of accounts is quite common among family members or roommates, the design of efficient recommender schemes for shared accounts has raised much attention recently. Generally speaking, after each login, it is essential for a recommender system to identify the current user behind and leverage this information to make recommendations. One naive approach is first to identify the identity of the current user and then make recommendations. However, this two-stage based approach may not achieve satisfactory performance. The key is that the recommended items favoring identifying users in the first stage may not be interesting to the users, which can deplete the user's patience quickly and cause early termination of users. To address the problem, we propose a novel recommendation scheme that makes a tradeoff between recommending discriminating items (helpful for identifying the user) and recommending interesting ones to the user (helpful for increasing the number of clicks). Under this scheme, we develop a patience model to capture the user's dynamic patience level during the recommendation process. Moreover, considering the increasing popularity of mobile devices, we also incorporate mobile sensor data (i.e., angle, accelerometer, gyroscope, etc.) into our approach to further improve the performance of the system. We implemented the above system in an App on mobile devices and carried out extensive experiments. The results demonstrate that our proposed scheme significantly outperforms the existing state-of-the-art approaches. [ABSTRACT FROM AUTHOR]

Genomic Epidemiology (genEpi) is a branch of public health that uses many different data types including tabular, network, genomic, and geographic, to identify and contain outbreaks of deadly diseases. Due to the volume and variety of data, it is challenging for genEpi domain experts to conduct data reconnaissance; that is, have an overview of the data they have and make assessments toward its quality, completeness, and suitability. We present an algorithm for data reconnaissance through automatic visualization recommendation, GEViTRec. Our approach handles a broad variety of dataset types and automatically generates visually coherent combinations of charts, in contrast to existing systems that primarily focus on singleton visual encodings of tabular datasets. We automatically detect linkages across multiple input datasets by analyzing non-numeric attribute fields, creating a data source graph within which we analyze and rank paths. For each high-ranking path, we specify chart combinations with positional and color alignments between shared fields, using a gradual binding approach to transform initial partial specifications of singleton charts to complete specifications that are aligned and oriented consistently. A novel aspect of our approach is its combination of domain-agnostic elements with domain-specific information that is captured through a domain-specific visualization prevalence design space. Our implementation is applied to both synthetic data and real Ebola outbreak data. We compare GEViTRec's output to what previous visualization recommendation systems would generate, and to manually crafted visualizations used by practitioners. We conducted formative evaluations with ten genEpi experts to assess the relevance and interpretability of our results. Code, Data, and Study Materials Availability: https://github.com/amcrisan/GEVitRec. [ABSTRACT FROM AUTHOR]

Visualization recommendation (VisRec) systems provide users with suggestions for potentially interesting and useful next steps during exploratory data analysis. These recommendations are typically organized into categories based on their analytical actions, i.e., operations employed to transition from the current exploration state to a recommended visualization. However, despite the emergence of a plethora of VisRec systems in recent work, the utility of the categories employed by these systems in analytical workflows has not been systematically investigated. Our article explores the efficacy of recommendation categories by formalizing a taxonomy of common categories and developing a system, Frontier, that implements these categories. Using Frontier, we evaluate workflow strategies adopted by users and how categories influence those strategies. Participants found recommendations that add attributes to enhance the current visualization and recommendations that filter to sub-populations to be comparatively most useful during data exploration. Our findings pave the way for next-generation VisRec systems that are adaptive and personalized via carefully chosen, effective recommendation categories. [ABSTRACT FROM AUTHOR]

Strategic traffic management is crucial for combating traffic congestion at the macroscopic level. However, such a field is still relatively unexplored, particularly for microscopic control objects, such as intersections and coordinated intersection groups. This article proposes a human-in-the-loop recommendation system for strategic urban traffic management, which follows an agent-based structure. A regional agent dispatcher is defined to assign agents for operation whenever “operation on-demand” is required. Such a requirement is identified by a daily-dependent operational mode on strategic traffic operations at a control object level. The strategic management scheme for each control object is guided by a strategic agent (customized), which is essentially a deep recommender model with a specific architecture. By featuring the multiagent design, a customized operational scheme can be generated at the intersection level, which instructs the corresponding controller to take specific operations. The utility of the recommendation system is demonstrated via a case study using real-world traffic data. In both offline and online evaluations, the system performs consistently at traffic operational recommendations in different scenarios and has the potential to provide more reasonable traffic operational strategies than a human-operated system. [ABSTRACT FROM AUTHOR]

This paper introduces versatile filters to construct efficient convolutional neural networks that are widely used in various visual recognition tasks. Considering the demands of efficient deep learning techniques running on cost-effective hardware, a number of methods have been developed to learn compact neural networks. Most of these works aim to slim down filters in different ways, e.g., investigating small, sparse or quantized filters. In contrast, we treat filters from an additive perspective. A series of secondary filters can be derived from a primary filter with the help of binary masks. These secondary filters all inherit in the primary filter without occupying more storage, but once been unfolded in computation they could significantly enhance the capability of the filter by integrating information extracted from different receptive fields. Besides spatial versatile filters, we additionally investigate versatile filters from the channel perspective. Binary masks can be further customized for different primary filters under orthogonal constraints. We conduct theoretical analysis on network complexity and an efficient convolution scheme is introduced. Experimental results on benchmark datasets and neural networks demonstrate that our versatile filters are able to achieve comparable accuracy as that of original filters, but require less memory and computation cost. [ABSTRACT FROM AUTHOR]

Click-through rate (CTR) prediction, whose goal is to predict the probability of the user to click on an item, has become increasingly significant in the recommender systems. Recently, some deep learning models with the ability to automatically extract the user interest from his/her behaviors have achieved great success. In these work, the attention mechanism is used to select the user interested items in historical behaviors, improving the performance of the CTR predictor. Normally, these attentive modules can be jointly trained with the base predictor by using gradient descents. In this paper, we regard user interest modeling as a feature selection problem, which we call user interest selection. For such a problem, we propose a novel approach under the framework of the wrapper method, which is named Meta-Wrapper. More specifically, we use a differentiable module as our wrapping operator and then recast its learning problem as a continuous bilevel optimization. Moreover, we use a meta-learning algorithm to solve the optimization and theoretically prove its convergence. Meanwhile, we also provide theoretical analysis to show that our proposed method 1) efficiencies the wrapper-based feature selection, and 2) achieves better resistance to overfitting. Finally, extensive experiments on three public datasets manifest the superiority of our method in boosting the performance of CTR prediction. [ABSTRACT FROM AUTHOR]

The Area under the ROC curve (AUC) is a well-known ranking metric for problems such as imbalanced learning and recommender systems. The vast majority of existing AUC-optimization-based machine learning methods only focus on binary-class cases, while leaving the multiclass cases unconsidered. In this paper, we start an early trial to consider the problem of learning multiclass scoring functions via optimizing multiclass AUC metrics. Our foundation is based on the M metric, which is a well-known multiclass extension of AUC. We first pay a revisit to this metric, showing that it could eliminate the imbalance issue from the minority class pairs. Motivated by this, we propose an empirical surrogate risk minimization framework to approximately optimize the M metric. Theoretically, we show that: (i) optimizing most of the popular differentiable surrogate losses suffices to reach the Bayes optimal scoring function asymptotically; (ii) the training framework enjoys an imbalance-aware generalization error bound, which pays more attention to the bottleneck samples of minority classes compared with the traditional $O(\sqrt{1/N})$ O (1 / N) result. Practically, to deal with the low scalability of the computational operations, we propose acceleration methods for three popular surrogate loss functions, including the exponential loss, squared loss, and hinge loss, to speed up loss and gradient evaluations. Finally, experimental results on 11 real-world datasets demonstrate the effectiveness of our proposed framework. The code is now available at https://github.com/joshuaas/Learning-with-Multiclass-AUC-Theory-and-Algorithms. [ABSTRACT FROM AUTHOR]

Near-data processing (NDP) architecture is promised to break the bottleneck of data movement in many scenarios (e.g., databases and recommendation systems), which limits the efficiency of data processing. Different from traditional SSD, NDP-based SSD not only needs to handle normal I/Os (e.g., read and write), but also needs to handle NDP requests that contain data processing operations. NDP and normal I/O requests share some function units of NDP-based SSD, such as flash chips and embedded processors. However, existing works ignore the resource competition between normal I/Os and NDP requests, which drastically degrades the performance. In this article, we propose a novel scheduling technique called Horae, which can efficiently schedule hybrid NDP-normal I/O requests in NDP-based SSD to improve performance. Horae exploits the critical paths on critical resources to maximize the parallelism of multiple stages of requests. The experimental results on typical workloads show that Horae can significantly improve the performance of hybrid NDP-normal I/O requests over the state-of-the-art scheduling algorithms of NDP-based SSDs. [ABSTRACT FROM AUTHOR]

Contribution: Instructors are leveraging open-source software (OSS) as a way to experience authentic examples of software problems with their students. Recommender engines might assist students in selecting the right project based on metrics mined from project repositories (e.g., GitHub). This vision is realized through GitMate, a GitHub-based recommender for supporting students in their OSS selection. Background: Contributing to OSS is a valuable way to immerse students into the realities of software development. When it comes to OSS selection, self-selection seems to be the most engaging alternative. Yet, students lack the time (and skills) to analyze project facets and draw comparisons among OSS projects. Research Questions: How can students be assisted to select a good OSS project to contribute to? Specifically, how would a recommender system might help? The envisioned intervention should be useful not only in finding the right project but also challenging students’ initial selections with other alternatives, spurring reflection. Methodology: The aim is to act upon a dependent variable (mind changing in project selection) through an independent variable (project comparison). This is achieved through GitMate, a recommender system on top of GitHub. Its search facilities are used for students to locate three projects at their wish. Next, GitMate recommends similar projects based on the project facets (e.g., number of committers, commits, and stars), mined from GitHub. Pondering the importance of distinct facets, students can now tradeoff different projects. The experiment checks whether students change their first choice. Findings: The results indicate that GitMate helps students compare GitHub projects to the extent of making them change their first choice. Nearly, 80% of the students changed at least one project as a result of using GitMate. This seems to suggest GitMate being effective on its goal: facet-based comparison thinking during OSS selection. [ABSTRACT FROM AUTHOR]

Collaborative filtering (CF) is a classical model for recommender systems. Though neural network-based CF models have been shown to perform well in some cases, they still suffer from the sparsity and cold-start issues. To address these issues, we propose a novel neural network-based CF model, termed the cross-domain explicit–implicit-mixed CF neural network (CEICFNet). The proposed model utilizes deep neural networks to learn latent factors not only from the explicit ratings and the implicit interactions but also in a cross-domain manner. In particular, domain-shared multilayer perception (MLP) networks are designed to learn the user rating latent factors and the user interaction latent factors from the explicit ratings and the implicit interactions, respectively, which act as bridges for transferring knowledge across domains. On the other hand, domain-specific MLP networks are designed to learn the item rating latent factors and the item interaction latent factors from the explicit ratings and the implicit interactions, respectively. Then, in each domain, based on the user rating latent factors and the item rating latent factors, the rating predictive representation for each user–item pair can be learned by an MLP. Similarly, the interaction predictive representation for each user–item pair can be learned. For integrating the explicit ratings and the implicit interactions, in each domain, a fully connected layer is used to automatically assign different weights to the rating predictive representations and the interaction predictive representations, based on which the final interaction probability can be generated. Extensive experiments are conducted on five datasets and the results have confirmed the effectiveness of our model. [ABSTRACT FROM AUTHOR]

Over two decades, context awareness has been incorporated into recommender systems in order to provide, not only the top-rated items to consumers but also the ones that are suitable to the user context. As a class of context-aware systems, context-aware service recommendation (CASR) aims to bind high-quality services to users, while taking into account their context requirements, including invocation time, location, social profiles, connectivity, and so on. However, current CASR approaches are not scalable with the huge amount of service data (QoS and context information, users reviews and feedbacks). In addition, they lack a rich representation of contextual information, as they adopt a simple matrix view. Moreover, current CASR approaches adopt the traditional user-service relation and they do not allow for multi-relational interactions between users and services in different contexts. To offer a scalable and context-sensitive service recommendation with great analysis and learning capabilities, we provide a rich and multi-relational representation of the CASR knowledge, based on the concept of knowledge graph. The constructed context-aware service knowledge graph (C-SKG) is, then, transformed into a low-dimensional vector space to facilitate its processing. For this purpose, we adopt Dilated Recurrent Neural Networks to propose a context-aware knowledge graph embedding, based on the principles of first-order and subgraph-aware proximity. Finally, a recommendation algorithm is defined to deliver the top-rated services according to the target user's context. Experiments have proved the accuracy and scalability of our solution, compared to state-of-the-art CASR approaches. [ABSTRACT FROM AUTHOR]

With the boosting of neural networks, recommendation methods become significantly improved by their powerful ability of prediction and inference. Existing neural-network based recommender systems (NN-RSs) usually first employ matrix embedding (ME) as a pre-process to learn users’ and items’ representations (latent vectors), then input these representations to a specific modified neural network framework to make accurate Top-k recommendations. Obviously, the performance of ME has a significant effect on RS models. However, most NN-RSs focus on accuracy by building representations from the direct user-item interactions (e.g., user-item rating matrix), while ignoring the underlying relatedness between users and items (e.g., users who rate the same ratings for the same items should be embedded into similar representations), which is an ideological disadvantage. On the other hand, ME models directly employ inner products as a default loss function metric that cannot project users and items into a proper latent space, which is a methodological disadvantage. In this paper, we propose a supervised collaborative representation learning model - Magnetic Metric Learning (MML) - to map users and items into a unified latent vector space, enhancing the representation learning for NN-RSs. First, MML utilizes dual triplets to model not only the observed relationships between users and items, but also the underlying relationships between users as well as items to overcome the ideological disadvantage. Specifically, a modified metric-based dual loss function is proposed in MML to gather similar entities and disperse the dissimilar ones. With MML, we can easily compare all the relationships (user to user, item to item, user to item) according to the weighted metric, which overcomes the methodological disadvantage. We conduct extensive experiments on four real-world datasets with large item space. The results demonstrate that MML can learn a proper unified latent space for representations from the user-item matrix with high accuracy and effectiveness, and lead to a performance gain over the state-of-the-art RS models by an average of 17 percent. [ABSTRACT FROM AUTHOR]

Human behaviors in recommendation systems are driven by many high-level, complex, and evolving intentions behind their decision making processes. In order to achieve better performance, it is important for recommendation systems to be aware of user intentions besides considering the historical interaction behaviors. However, user intentions are seldom fully or easily observed in practice, so that the existing works are incapable of fully tracking and modeling user intentions, not to mention using them effectively into recommendation. In this paper, we present the Intention-Aware Sequential Recommendation (ISRec) method, for capturing the underlying intentions of each user that may lead to her next consumption behavior and improving recommendation performance. Specifically, we first extract the intentions of the target user from sequential contexts, then take complex intent transition into account through the message-passing mechanism on an intention graph, and finally obtain the future intentions of this target user from inference on the intention graph. The sequential recommendation for a user will be made based on the predicted user intentions, offering more transparent and explainable intermediate results for each recommendation. Extensive experiments on various real-world datasets demonstrate the superiority of our method against several state-of-the-art baselines in sequential recommendation in terms of different metrics. [ABSTRACT FROM AUTHOR]

Recommender systems (RS) are increasingly leveraging the power of graphs to enhance accuracy. However, we stipulate that existing methods don’t take into consideration the inherent behavior of communities and the interaction between all the sub-groups of the network. In this work, we develop a Deep Graph-based Collaborative Filtering recommender system (DGCF), which incorporates the concept of community profiling and leverages the power of Graph Neural Networks. DGCF utilizes multiple graphs to exploit all types of information from the different user interactions. It extracts the overlapping communities from the homophily user-user graph and also integrates the high-order information from the user-item bipartite graph. We conduct experiments and evaluate the DGCF on the MovieLens datasets (ML-100K and ML-1M), and Douban dataset. Our experiments reveal significant improvements over a number of the latest deep learning models for recommender systems. Results also support that DGCF has the potential to render better recommendations as it extracts deep relationships using the community structure.

In the airline industry, the Revenue and Pricing teams generally spend a considerable amount of time analysing and interpreting the actions of their competitors. Most of the time the analysts have to use their analytical skills to create ad-hoc methods to interpret or find patterns in the fares. In this field, it is key to automate the process, avoid human errors, and add new features that provide accurate fares. Considering this, a gene expression programming algorithm is proposed to carry out this process, returning an interpretable rule set which acts as a recommender system to ease the daunting process done by the pricing teams manually. The proposal was applied to a real scenario with the information provided by the Air Canada airline for five months in Canadian markets. The experimental analysis revealed, by means of non-parametric statistical tests, that the proposed gene expression programming algorithm was key to getting the appropriate features and, hence, accurate and highly interpretable results. The proposal obtained extremely accurate results (around 96% in both accuracy and F1 measure) with a reduction of around 50% in the rule set in many cases.

The cold-start problem is one of the main challenges in recommender systems and specifically in collaborative filtering methods. Such methods, albeit effective, typically can not handle new items or users that do not have any prior interaction activity in the system. In this paper, we propose a novel two-step approach to address the cold-start problem. First, we view the user-item interactions in a positive unlabeled (PU) learning setting and reconstruct the interaction matrix between users and warm items, detecting missing links and recommending warm items to existing users. Second, an inductive multi-target regressor is trained on this reconstructed interaction matrix and subsequently predicts interactions for new items that enter the system. To the best of our knowledge, this is the first time that such a two-step PU learning method is proposed to address the cold-start problem in recommender systems. To evaluate the proposed approach, we employed four benchmark datasets from movie and news recommendation domains with explicit and implicit feedback. We compared our method against three other competitor approaches that address the cold-start problem and showed that our proposed method significantly outperforms them, achieving in a case an increase of 16.9% in terms of NDCG.

An Explainable Recommendation system has become more prominent because it could convince the user to trust the prediction. However, the existing Explainable Recommendation systems face one of these two limitations. The first limitation is the utilization of a pre-defined feature set to represent the users/items’ preferences. By representing the preferences with a pre-defined feature set, the uniqueness of each user will be limited by the characteristics of the pre-defined set. For the other limitation, the user/item preference representation in existing models is constructed from the entire past interactions. This can mislead the model because the user’s preferences change over the interaction’s characteristics. Therefore, the entire past interactions might not be able to represent the user’s preferences for the target interaction. In this paper, we proposed an Explainable Recommendation system, which represents the user’s preferences with a personalized feature set and is able to adapt the target user/item preference representation to suit the target interaction’s characteristics. By representing the users/items’ preferences with a personalized feature set, the proposed method is able to give the score of importance in the rating contribution to each feature and uses these scores to assume the reasons for the user’s decision. To make the representation suitable to the target interaction’s characteristics, the proposed model utilizes only the target interaction’s information to represent the preferences of users/items. The proposed method also is not relying on the entire past interactions to avoid the bias of dominant features in the past interactions. For the experiment, the proposed method consistently outperforms all baselines on three benchmark datasets. We also analyze the explanation resulting from the proposed model to support our assumptions that the preferences of the user are unique and dynamically change over the interaction’s characteristics.

Recommender Systems have become one of the most important tools for streaming and marketplace systems in recent years. Their increased use has revealed clear bias and unfairness against minorities and underrepresented groups. This paper seeks the origin of these biases and unfairness. To this end, it analyzes the demographic characteristics of a gold standard dataset and its prediction performance when used in a multitude of Recommender Systems. In addition, this paper proposes Soft Matrix Factorization (SoftMF), which tries to balance the predictions of different types of users to reduce the present inequality. The experimental results show that those biases and unfairness are not introduced by the different recommendation models and that they come from the socio-psychological and demographic characteristics of the used dataset.