Recommendation Problem Research Articles

Enhancing federated learning-based social recommendations with graph attention networks

Social recommendations are crucial in information services, particularly for mitigating information overload on the Internet. Graph Attention Networks (GATs) have emerged as an effective solution for the social recommendation problem, harnessing the benefits of representation learning from user-item and user-user graphs. However, traditional GAT-based social recommendation models often necessitate the centralized storage and aggregation of all the users’ social connections and item interactions. This centralization raises significant privacy concerns. To address these issues, we propose a federated learning-based social recommendation system using GATs. Specifically, we first design a structured-enhanced graph clustering algorithm to obtain soft clustering assignments for items. Based on these assignments, the similarity between pairwise items and their corresponding correlations are computed. Second, we employ graph attention networks to infer the local heterogeneity of social relationships, user-item interactions, and item relationships within the federated learning framework for each client. Using the soft clustering assignments, we derive hard clustering indicators and sample pseudo items. Local gradients of embedding and model parameters are then computed and uploaded to the central server for aggregation under the local differential privacy mechanism. Experimental results over real-world data sets indicate that the proposed model is more effective in privacy protection while maintaining acceptable performance compared with the central learning framework. Additionally, the proposed model is more stable in addressing the cold-start issue for items than state-of-the-art federated learning-based social recommendation models.

DualCFGL: dual-channel fusion global and local features for sequential recommendation

Sequential recommendation systems capture the dynamic interests of users and predict their future preferences. A noteworthy problem in sequential recommendation is coping with the intrinsic changes of user interests. The sequence of user interactions is generated by more than a single and stable global preference, users may have interest drift that occur in a short period of time. We call this short-term interest drift as the local preference of users, which is often a key factor affecting the final choice of users. However, existing methods have limitations in observing local preferences, which leads to an incomplete consideration of the local preferences. Moreover, using a single model to represent global–local preferences obscure the distinct features of each, limiting the potential synergistic benefits. To alleviate the above limitations, we propose a novel model with a dual-channel structure to monitor both global and local preferences and ensure they complement each other. The model extracts the global preferences of users with a bidirectional Transformer using random masking and a sliding window, and extracts the local preferences with a patch-based stacked bottleneck residual convolution. To enable the model to consider both the global and local preferences of users, we design an adaptive orthogonal fusion module, which effectively fuses the two preferences and enables the two feature types to complement and enhance each other. We integrate the fused user preferences with a knowledge distillation method that further improves the model’s expressive ability. We conduct extensive experiments on three widely used datasets, and the results show that our model outperforms current state-of-the-art models.

Applying Recommender Systems to Predict Personalized Film Age Ratings for Parents

A motion picture content rating system categorizes a film based on its appropriateness for various audiences, considering factors such as portrayals of sex, violence, substance abuse, profanity, and other elements typically considered unsuitable for children or adolescents. This rating is usually coupled with a minimum desired age that the film is suitable for. In this work, we apply recommender systems to predict personalized film age ratings for parents. According to the proposed methodology, we reduce the personalized film age prediction problem to the classic item recommendation problem by applying a recommender system for each age film category. The recommender systems generate recommendations for each film age category. Finally, these recommendations are combined to provide the final age recommendation for the parent (user). The proposed methodology was applied to state-of-the-art recommender systems. In addition, we used them as baselines for comparing the direct application of a recommender system to the age prediction problem. This was achieved by treating each film as an item and assigning the given age as its rating. The experimental results highlight the efficiency of the proposed system when applied to a well-known real-world dataset.

Development of web based data-driven recommendation system for house rental via hierarchical fuzzy axiomatic design

PurposeIn the real estate business, identifying the ideal property for a user poses a difficult task due to the many factors involved in the decision-making process. Moreover, users often struggle to find platforms that facilitate effective communication of their preferences. To tackle this issue, a web-based data-driven recommendation system has been devised for the real estate business.Design/methodology/approachThe process of identifying the most suitable rental property for a user hinges greatly on how the user prioritizes each criterion and the analysis of unstructured data. In this research, a novel recommendation system for house rentals is developed by utilizing the Weighted Hierarchical Fuzzy Axiomatic Design (WFAD) approach. Techniques for extracting pertinent information from unstructured house descriptions are employed. The user’s preferences are captured through an interactive web application equipped with a map feature to highlight key locations.FindingsData on various available rental properties are gathered using web scraping techniques. The efficacy of the proposed rental house recommendation system is demonstrated through multiple case studies. It is observed that the developed system provides more informed and reliable decisions.Originality/valueFirst time in the related literature, we applied the weighted fuzzy axiomatic design procedure (WFAD) to the product recommendation problem and developed a comprehensive web-based system for recommending rental houses based on it in the real estate business.

A Research on Recommendation Method of Kill Chain Based on Graph Theory Algorithm

Abstract The present study introduces an optimization method based on graph theory to address the kill chain recommendation problem in rapidly changing modern battlefield environments. This method transforms the kill chain selection problem into a shortest path problem with time constraints and designs an efficient optimization process using branch-and-bound algorithm. This study constructs an evaluation model for kill chain efficacy and time utility to ensure that the recommended kill chain achieves optimal efficacy within the target’s active strike time. The efficacy and efficiency of the proposed algorithm in handling multi-target scenarios have been substantiated through the analysis of data procured from the “2020 Joint Operations Smart Game Challenge”, reducing the optimization time when compared to traditional methods. This study provides a new perspective and technical assistance to enhance the intelligence and automation of operational command decision-making processes.

A model to address the cold-start in peer recommendation by using k-means clustering and sentence embedding

In academia, research collaboration plays a vital role in enhancing the research quality and enriching the academic profile of the authors. Recommending appropriate collaborators from a vast scholarly database, particularly for newcomers, poses a challenging cold-start problem. This study addresses a cold-start problem in peer recommendation, considering a dynamic coauthorship graph as a network structure of academic collaborators. As the coauthorship graph is quite large and complex, an efficient indexing method is essential for speeding up the initial search of similar coauthors. The study introduces an efficient Global Inverted List (GIL) for indexing research areas and active authors in the coauthorship network. An attribute-based search and filtering mechanism is proposed to identify relevant collaborators, followed by the application of k-means clustering and doc2vec metrics to rank and select top recommendations. A cold user is associated with attributes that identify coauthors with similar research interests. For each attribute of the cold user, the model searches the associated authors from the GIL. Further, two filtering approaches are applied to refine the retrieved author list. The first ensures that the authors have a significant presence in the specified research areas, whereas the second one helps avoid recommending authors with only superficial connections to the cold user. The model creates a feature matrix of filtered authors using the publication features of authors. The k-means clustering applied to the feature matrix generates k clusters, among which the model chooses only those with seed nodes i.e. the clusters which are having seed nodes are selected for further process. Selected clusters are ranked using doc2vec metrics, with the top-ranked cluster providing the final recommendation. The model recommends the top L members of the selected cluster, where L is the length of the recommendations provided to the new user. Our extensive experiments show the efficacy of the proposed model.

An enhanced genetic algorithm solution for itinerary recommendation considering various constraints.

This paper addresses the challenging task of itinerary recommendation for tourists and proposes an approach for suggesting efficient optimal itineraries in Istanbul, based on constraints. The paper presents an enhanced version of the genetic algorithm (GA), which aims to optimize the itineraries considering various constraints and preferences of the tourists. The improvement of the GA involved suggesting a customized fitness function tailored to address the complexities of the tourism problem, considering factors such as distance, time, cost, tourists' budget, and their desired activities and attractions. Additionally, we proposed a new crossover method, named "Copy Order Crossover" and we modified the tournament selection method beside enhancing the implementation of the swap mutation method for greater efficiency and adaptability. The enhanced GA is evaluated on the Burma dataset taken from TSPLIB, and our constructed Istanbul dataset, achieving significant enhancement rates in GA (43.89% for Istanbul, and 56.60% for Burma). This paper provides a detailed account of the proposed approach, its implementation, and the evaluation conducted. The experimental results conclusively demonstrated the superiority of the proposed approach over alternative methods in terms of time, efficiency, and accuracy. This paper finishes with an outlook with a detailed potential approach to overcome itinerary recommendation problem limitations.

Recommendation-Based Trust Evaluation Model for the Internet of Underwater Things

The Internet of Underwater Things (IoUT) represents an emerging and innovative field with the potential to revolutionize underwater exploration and monitoring. Despite its promise, IoUT faces significant challenges related to reliability and security, which hinder its development and deployment. A particularly critical issue is the establishment of trustworthy communication networks, necessitating the adaptation and enhancement of existing models from terrestrial and marine systems to address the specific requirements of IoUT. This work explores the problem of dishonest recommendations within trust modelling systems, a critical issue that undermines the integrity of trust modelling in IoUT networks. The unique environmental and operational constraints of IoUT exacerbate the severity of this issue, making current detection methods insufficient. To address this issue, a recommendation evaluation method that leverages both filtering and weighting strategies is proposed to enhance the detection of dishonest recommendations. The model introduces a filtering technique that combines outlier detection with deviation analysis to make initial decisions based on both majority outcomes and personal experiences. Additionally, a belief function is developed to weight received recommendations based on multiple criteria, including freshness, similarity, trustworthiness, and the decay of trust over time. This multifaceted weighting strategy ensures that recommendations are evaluated from different perspectives to capture deceptive acts that exploit the complex nature of IoUT to the advantage of dishonest recommenders. To validate the proposed model, extensive comparative analyses with existing trust evaluation methods are conducted. Through a series of simulations, the efficacy of the model in capturing dishonest recommendation attacks and improving the accuracy rate of detecting more sophisticated attack scenarios is demonstrated. These results highlight the potential of the model to significantly enhance the trustworthiness of IoUT establishments.

Multi-Granularity and Multi-Interest Contrast-Enhanced Hypergraph Convolutional Networks for Session Recommendation

Aiming at the problems of the incomplete recommendation and sparsity of session data in session recommendation, a new multi-granularity and multi-interest contrast-enhanced hypergraph convolutional network (MGMI-CEHCN) model for session recommendation is proposed. The session data are modeled as a heterogeneous hypergraph, the information is embedded by the hypergraph using two granularities of item price and category, and then the information between different granularities is fused, while the final item embedding is obtained through multi-layer convolution. Finally, an interest perceptron is used to detect multiple potential interests for each item, and a decentralized interest extraction network based on a gated recurrent unit (GRU) is used to integrate the user’s final interests and obtain a global session representation through a soft attention mechanism; a local session representation is generated with the help of a weighted line graph convolutional network. A further joint contrast enhancement strategy is used to maximize the mutual information between the global session representation and local session representation, to improve the recommendation performance. Experiments on several real datasets showed that the recommendation performance of the MGMI-CEHCN model outperformed the current mainstream models. On Cosmetics, the P@20 reached up to 55.25% and M@20 reached up to 38.26%, improvements of 3.06% and 3.09%, respectively; on Diginetica-buy, the P@20 reached up to 65.60% and M@20 reached up to 27.47%, improvements of 2.47% and 6.64%, respectively, which proved the validity of the model.

Quantum contextual bandits and recommender systems for quantum data

We study a recommender system for quantum data using the linear contextual bandit framework. In each round, a learner receives an observable (the context) and has to recommend from a finite set of unknown quantum states (the actions) which one to measure. The learner has the goal of maximizing the reward in each round, that is the outcome of the measurement on the unknown state. Using this model, we formulate the low energy quantum state recommendation problem where the context is a Hamiltonian and the goal is to recommend the state with the lowest energy. For this task, we study two families of contexts: the Ising model and a generalized cluster model. We observe that if we interpret the actions as different phases of the models, then the recommendation is done by classifying the correct phase of the given Hamiltonian, and the strategy can be interpreted as an online quantum phase classifier.

Service recommendation in JointCloud environments: An efficient regret theory-based Qos-aware approach

With the proliferation of data-intensive applications, there arises an urgent demand for a substantial amount of cloud services to meet their requirements for data analysis. This globalized yet cooperative business landscape necessitates new cooperative models across the world. JointCloud, as a novel cross-cloud cooperation computing model, takes the first step towards establishing an evolving cloud ecosystem where all cloud service providers could collaboratively serve globalized computation needs. The collaboration among various cloud service providers enhances both the availability and Quality of Services(QoS) of cloud services, enabling a cloud service provider to concurrently serve users with differentiated QoS requirements. This unique characteristic further complicates the problems of QoS-aware service recommendations, rendering conventional approaches obsolete and inefficient. Thus, there is an urgent need to improve the efficiency and effectiveness of the service recommendation method, which is of vital importance for the JointCloud environment. In this paper, we present a two-stage efficient regret theory-based service recommendation method for the JointCloud environment. In the first stage of our proposed method, we cluster the cloud service providers to reduce the choice space to improve the efficiency of cloud service recommendations. In the second stage, we meticulously identify the most appropriate services within one cluster. To enhance the overall rationality of service recommendation, we introduce a subjective and objective combined weighting method and a regret theory based ranking method. Extensive experimental results demonstrate that our approach can facilitate fast and accurate service recommendations.

Adaptive Preference Measurement with Unstructured Data

Many products are most meaningfully described using unstructured data such as text or images. Unstructured data are also common in e-commerce, in which products are often described by photos and text but not with standardized sets of attributes. Whereas much is known about how to efficiently measure consumer preferences when products can be meaningfully described by structured attributes, there is scant research on doing the same for unstructured data. This paper introduces a real-time, adaptive survey design framework for measuring preferences over unstructured data, leveraging Bayesian optimization. By adaptively choosing items to display based on uncertainty around a nonparametric utility model, the proposed method maximizes information gain per question, enabling quick estimation of individual-level preferences. The approach operates on embeddings of the unstructured data, thereby eliminating the requirement for manual coding of product attributes. We apply the method to measuring preferences over clothing and highlight its potential for both the general task of marketing research and the specific task of designing customer onboarding surveys to mitigate the cold-start recommendation problem. We also develop methods for interpreting the nonparametric utility functions, which allow us to reconstruct consumer valuations of discrete attributes, even for attributes that were not considered or available a priori. This paper was accepted by Duncan Simester, marketing. Fundings: Funding for this project was provided by Analytics at Wharton, the Wharton Behavioral Lab, and the Wharton Dean’s Fund. The author also thanks the Govil Family for financial support. Supplemental Material: The online appendix and data files are available at https://doi.org/10.1287/mnsc.2023.03775 .

Recommending for a Multi-Sided Marketplace: A Multi-Objective Hierarchical Approach

We propose a multi-objective hierarchical recommender that solves the multi-objective recommendation problem in a hierarchical setting and was deployed on Uber Eats.

TSDRL: A three-stage deep reinforcement learning method for reliable collaboration of manufacturing service towards mass personalized production

With the development of intelligent manufacturing driven by new-generation information technology, manufacturing service collaborative chains (MSCC) based on Industrial Internet Platforms have become the main way to cope with mass personalized production. To adapt the dynamics and variability of personalized demands, manufacturing services are required to be frequently reorganized. This leads to dynamic changes in manufacturing service availability and ambiguous reliable collaboration mechanisms, making it challenging for existing methods to consistently maintain stable MSCC. Therefore, this paper proposes a three-stage deep reinforcement learning method (TSDRL) to improve the reliability of MSCC while guaranteeing the satisfaction of personalized demands. The method comprehensively considers the dynamic changes of manufacturing services and service reliability, and solves the reliable MSCC recommendation problem. Specifically, in stage one, high-quality services are selected through DRL state. In stage two, the improved NSGA-III algorithm is used to optimize the MSCC. In stage three, the BP+GA algorithm is used to mine user preferences and recommend high-satisfaction MSCC. By leveraging the exploratory process of DRL, cyclic optimization of MSCC is performed. Through comprehensive analysis and experimental research, it is shown that the method of this paper is robust and practical in dynamic environment.

A recommendation approach of scientific non-patent literature on the basis of heterogeneous information network

In the procedure of exploring science-technology linkages, non-patent literature (NPL) in patents, particularly scientific NPL, is considered to signal the relatedness between the developed technology and the cited science. However, many prior art search tools may not be powered with the cross-collection recommendation technique, or have limited cross-collection recommendation capabilities. In this paper, we present an approach to recommend scientific NPL for a focal patent on the basis of heterogeneous information network. This study views this cross-collection recommendation problem as a link prediction problem on the basis of meta-path counting approach. Extensive experiments on DrugBank dataset in the pharmaceutical field indicate that our approach is feasible and effective. This work provides a novel perspective on scientific NPL recommendation for a focal patent and opens up further possibilities for the linkages between science and technology. Nevertheless, more experiments in other fields are required to verify the recommended effects of the approach proposed in this study.

X-Model4Rec: An Extensible Recommender Model Based on the User’s Dynamic Taste Profile

Several approaches have been proposed to obtain successful models to solve complex next-item recommendation problem in non-prohibitive computational time, such as by using heuristics, designing architectures, and applying information filtering techniques. In the current technological scenario of artificial intelligence, sequential recommender systems have been gaining attention and they are a highly demanding research area, especially using deep learning in their development. Our research focuses on an efficient and practical model for managing sequential session-based recommendations of specific products for users using the wine and movie domains as case studies. Through an innovative recommender model called X-Model4Rec – eXtensible Model for Recommendation, we explore the user's dynamic taste profile using architectures with transformer and multi-head attention mechanisms to solve the next-item recommendation problem. The performance of the proposed model is compared to that of classical and baseline recommender models on two real-world datasets of wines and movies, and the results are better for most of the evaluation metrics.

Book Classification and Recommendation for University Libraries using Grey Correlation and Bayesian Probability

In today’s information era, collaborative filtering algorithms are widely used and their distinct knowledge discovery techniques can effectively address numerous issues. However, conventional collaborative filtering algorithms encounter cold-start and data sparsity issues, which restrict their performance and accuracy. The study selected the multi-feature method to improve the traditional collaborative filtering algorithm, and introduced gray correlation calculation and Bayesian probability for user preference analysis. A learning resource recommendation model based on collaborative filtering was developed by comparing the target user’s characteristics with those of other users, calculating their similarity, selecting users with high similarity to the target user and forming a neighbor set. Using Bayesian probability and grey correlation to analyze user preferences in library systems can be well applied in book classification and recommendation problems in university libraries. The computing layer, which includes the collaborative filtering calculation stage and the group recommendation calculation stage, is the model’s main functional component. The smaller the value of mean absolute error, the higher the prediction accuracy of the model. The mean absolute error value of the multi-feature collaborative filtering algorithm was inferior to the traditional collaborative filtering algorithm, indicating that the classification accuracy of the former is higher than that of the latter. When the training set to test set ratio steadily became bigger, the mean absolute error value reached the lowest and smoothest point at 80%. In dataset A, the minimum mean absolute error values of multi-feature collaborative filtering and collaborative filtering were 0.765 and 0.809. Compared with traditional filtering algorithms, the mean absolute error value has decreased by 0.044. In dataset B, the mean absolute error values of multi-feature collaborative filtering and collaborative filtering were 0.796 and 0.836. Compared with traditional filtering algorithms, the mean absolute error value has decreased by 0.040. In dataset C, the minimum mean absolute error values of multi-feature collaborative filtering and collaborative filtering were 0.815 and 0.848. Compared with traditional filtering algorithms, the mean absolute error value has decreased by 0.033. When the accuracy was the highest; the mean absolute error value was the smallest at the grey correlation, which means that the technique improves the reliability of the recommendations compared with other methods. This means that the method has a positive impact on the accuracy of the recommendations compared to other methods. Grey correlation degree can comprehensively consider the interrelationships between multiple factors, handle uncertain and incomplete information, and explore potential user needs and behavior patterns. The implementation of the grey correlation degree has transformed the collaborative filtering algorithm into a group filtering algorithm, thereby enhancing its precision. The research on book classification and recommendation in university libraries, which enhances the group filtering algorithm, can address a range of issues such as improving classification accuracy, augmenting recommendation diversity, enhancing library management efficiency among others. This, in turn, enables more precise book recommendations to users.

Relation pruning and discriminative sampling over knowledge graph for long-tail recommendation

Long-tail recommendations have gained significant attention owing to their potential economic market. However, the scarcity of interaction data for long-tail users/items and the popularity bias present challenges in capturing high-quality embeddings for long-tail users/items. This predicament further exacerbates the long-tail recommendation problem, as current approaches tend to exhibit a bias towards making recommendations for short-head users/items, resulting in a detrimental cycle for long-tail recommendation. To this end, we propose a novel knowledge graph-based approach called LTailKG to improve long-tail recommendations. LTailKG leverages the semantic information in knowledge graph to produce high-quality embeddings and augmented samples for generating satisfactory long-tail recommendations. First, LTailKG parameterizes each node and relation as vector representations. Next, LTailKG presents a relation pruning-based graph contrastive learning operation to generate additional self-supervised signals for long-tail users/items, thereby producing high-quality embeddings for them. Furthermore, LTailKG introduces a knowledge graph-driven discriminative sampling operation to select augmented positive and negative samples from the uninteracted item set, which enables LTailKG to excel at not only identifying long-tail items that better align with the user's interests but also extracting the genuine preferences of long-tail users. Extensive experiments on real-world datasets demonstrate the superiority of LTailKG over state-of-the-art approaches in terms of long-tail recommendation.

Shopping trip recommendations: A novel deep learning-enhanced global planning approach

Brick-and-mortar shopping malls are embracing Artificial Intelligence (AI) technology and recommender systems to enhance the shopping experience and boost mall revenue. Echoing this trend, we formulate a new shopping trip recommendation problem, which aims to recommend a shopping trip (i.e., a list of stores) that matches customer preferences and has appropriate trip lengths. To solve this problem, we develop a novel deep learning-enhanced global planning (DeepGP) approach featuring three methodological novelties. First, we introduce a new shopping intensity term based on deep neural networks to capture the variation of trip lengths specific to different shopping contexts. Second, we innovatively formulate the learning and optimization objectives in a consistent form by balancing the shopping choice likelihood and the shopping intensity likelihood, thus resolving the inconsistency issue encountered by prior global planning methods. Third, to overcome the computational challenge caused by the nonlinear shopping intensity term, we design a new exact and efficient solution technique based on piecewise linear transformations. Using a real-world offline shopping dataset, we empirically demonstrate the superior performances of our approach compared to representative benchmarks in offering more accurate and relevant shopping trip recommendations. Through a simulation, we show the capacity of our approach to attract and balance customer traffic in practical deployments. Overall, our research highlights the efficacy of combining shopping choices and shopping intensity in a consistent learning and optimization framework for offline shopping trip recommendations.

Discrete Federated Multi-behavior Recommendation for Privacy-Preserving Heterogeneous One-Class Collaborative Filtering

Recently, federated recommendation has become a research hotspot mainly because of users’ awareness of privacy in data. As a recent and important recommendation problem, in heterogeneous one-class collaborative filtering (HOCCF), each user may involve of two different types of implicit feedback, that is, examinations and purchases. So far, privacy-preserving HOCCF has received relatively little attention. Existing federated recommendation works often overlook the fact that some privacy sensitive behaviors such as purchases should be collected to ensure the basic business imperatives in e-commerce for example. Hence, the user privacy constraints can and should be relaxed while deploying a recommendation system in real scenarios. In this article, we study the federated multi-behavior recommendation problem under the assumption that purchase behaviors can be collected. Moreover, there are two additional challenges that need to be addressed when deploying federated recommendation. One is the low storage capacity for users’ devices to store all the item vectors, and the other is the low computational power for users to participate in federated learning. To release the potential of privacy-preserving HOCCF, we propose a novel framework, named discrete federated multi-behavior recommendation (DFMR), which allows the collection of the business necessary behaviors (i.e., purchases) by the server. As to reduce the storage overhead, we use discrete hashing techniques, which can compress the parameters down to 1.56% of the real-valued parameters. To further improve the computation-efficiency, we design a memorization strategy in the cache updating module to accelerate the training process. Extensive experiments on four public datasets show the superiority of our DFMR in terms of both accuracy and efficiency.