Intra- and Inter-group Optimal Transport for User-Oriented Fairness in Recommender Systems

Cyber-physical systems (CPS) are often characterized as smart systems, which intelligently interact with other systems across information and physical interfaces. Cyber-physical systems (CPS) are fitting into modern society. CPS integrate multiple techniques, including distributed computing, communication and automatic control, to support variety of intelligent services and applications in many fields, such as transportation, healthcare, entertainment and city infrastructure. Recommender systems in CPS, which always provide information recommendations for users based on historical ratings collected from a single domain only, suffer from the data sparsity problem. Recently, several recommendation models have been proposed to transfer knowledge across multiple domains to alleviate the sparsity problem, which typically assumes that multiple domains share a latent common rating pattern. Recommender systems always results into faster and efficient options/choices as per user’s demands. This technology helps us in retrieval and access of different services like in Healthcare sector, Education, E-Commerce, etc. During the last few decades, with the rise of Youtube, Amazon, Netflix and many other such web services, recommender systems have taken more and more place in our lives. From e-commerce (suggest to buyers articles that could interest them) to online advertisement (suggest to users the right contents, matching their preferences), recommender systems are today unavoidable in our daily online journeys. In this paper, a proposed system architecture focusses on IoT applications utilizing recommender systems for offering number of services to users. Also, various Cyber Physical Recommender Systems are proposed to offer standard quality of service (QoS) for various IoT applications.

Recent studies reveal that recommender systems are vulnerable to data poisoning attack due to their openness nature. In data poisoning attack, the attacker typically recruits a group of controlled users to inject well-crafted user-item interaction data into the recommendation model's training set to modify the model parameters as desired. Thus, existing attack approaches usually require full access to the training data to infer items' characteristics and craft the fake interactions for controlled users. However, such attack approaches may not be feasible in practice due to the attacker's limited data collection capability and the restricted access to the training data, which sometimes are even perturbed by the privacy preserving mechanism of the service providers. Such design-reality gap may cause failure of attacks. In this paper, we fill the gap by proposing two novel adversarial attack approaches to handle the incompleteness and perturbations in user-item interaction data. First, we propose a bi-level optimization framework that incorporates a probabilistic generative model to find the users and items whose interaction data is sufficient and has not been significantly perturbed, and leverage these users and items' data to craft fake user-item interactions. Moreover, we reverse the learning process of recommendation models and develop a simple yet effective approach that can incorporate context-specific heuristic rules to handle data incompleteness and perturbations. Extensive experiments on two datasets against three representative recommendation models show that the proposed approaches can achieve better attack performance than existing approaches.

At present, it can be explained that the recommendation system is mostly designed for a specific recommendation model, and its scalability is weak. It is not enough for the emerging recommendation models, such as complex and mixed models with deep neural networks.Knowledge maps as a highly readable external knowledge carrier provide great possibilities for improving the ability of algorithms to interpret.In essence, the knowledge map is intended to describe the various entities or concepts and their relationships that exist in the real world, which constitute a huge semantic network diagram, nodes represent entities or concepts, and edges are composed of attributes or relationships.Based on the current research status, this paper analyzes and studies the concept of knowledge map, the combination of knowledge map and recommendation system, the object of interpretability and the application of knowledge map in interpretability model, combined with relevant recommendation models. The knowledge map, open up the relationship between the media, flexibly choose the most suitable medium according to the specific situation to recommend and explain the user.

In recommendation systems, the choice of loss function is critical since a good loss may significantly improve the model performance. However, manually designing a good loss is a big challenge due to the complexity of the problem. A large fraction of previous work focuses on handcrafted loss functions, which needs significant expertise and human effort. In this paper, inspired by the recent development of automated machine learning, we propose an automatic loss function generation framework, AutoLossGen, which is able to generate loss functions directly constructed from basic mathematical operators without prior knowledge on loss structure. More specifically, we develop a controller model driven by reinforcement learning to generate loss functions, and develop iterative and alternating optimization schedule to update the parameters of both the controller model and the recommender model. One challenge for automatic loss generation in recommender systems is the extreme sparsity of recommendation datasets, which leads to the sparse reward problem for loss generation and search. To solve the problem, we further develop a reward filtering mechanism for efficient and effective loss generation. Experimental results show that our framework manages to create tailored loss functions for different recommendation models and datasets, and the generated loss gives better recommendation performance than commonly used baseline losses. Besides, most of the generated losses are transferable, i.e., the loss generated based on one model and dataset also works well for another model or dataset. Source code of the work is available at https://github.com/rutgerswiselab/AutoLossGen.

Objective: To create an AI-powered recommendation system that is designed for IT professionals to help them choose the best software development approaches. Through the use of specified data parameters. Methods: The recommendation system will make use of machine learning algorithms and data analysis methods to examine team dynamics, project needs, and other variables. The technology will enable developers to improve the quality of products and speed up the development process by recommending suitable development methodologies. Data parameters considered for the development of the recommendation model fall into four categories: requirements, user involvement, development team, type of project, and risk associated with it. Findings: Existing recommendation systems developed by different researchers are applicable for only requirement elicitation and to recommend different phases of the development process, whereas systems that will help select development methodology are not available in the existing systems. Among the five machine learning algorithms applied in the recommender system building process, the DecisionTree Classifier and RandomForest Classifier exhibit superior performance, achieving 100% accuracy, while the Kneighbors Classifier indicates 94.74% accuracy. Novelty: This study of systems introduces a novel approach to software development methodology, a recommender system, which helps IT developers select the best appropriate development approach for the development of a software product or project based on the type of project to be built and other data parameters. Keywords: Agile, Development, Requirements, Methodology, User, Customer

Recommender systems play a pivotal role in mitigating information overload in various fields. Nonetheless, the inherent openness of these systems introduces vulnerabilities, allowing attackers to insert fake users into the system’s training data to skew the exposure of certain items, known as poisoning attacks. Adversarial training has emerged as a notable defense mechanism against such poisoning attacks within recommender systems. Existing adversarial training methods apply perturbations of the same magnitude across all users to enhance system robustness against attacks. Yet, in reality, we find that attacks often affect only a subset of users who are vulnerable. These perturbations of indiscriminate magnitude make it difficult to balance effective protection for vulnerable users without degrading recommendation quality for those who are not affected. To address this issue, our research delves into understanding user vulnerability. Considering that poisoning attacks pollute the training data, we note that the higher degree to which a recommender system fits users’ training data correlates with an increased likelihood of users incorporating attack information, indicating their vulnerability. Leveraging these insights, we introduce the Vulnerability-aware Adversarial Training (VAT), designed to defend against poisoning attacks in recommender systems. VAT employs a novel vulnerability-aware function to estimate users’ vulnerability based on the degree to which the system fits them. Guided by this estimation, VAT applies perturbations of adaptive magnitude to each user, not only reducing the success ratio of attacks but also preserving, and potentially enhancing, the quality of recommendations. Comprehensive experiments confirm VAT’s superior defensive capabilities across different recommendation models and against various types of attacks.

A survey of recommender systems with multi-objective optimization

Numerous recommender systems offer recommendations to users depending on their interests. Several applications including those in e-commerce, healthcare and markets have adopted recommendation systems. This paper's major goal is to demonstrate various difficulties with the methods utilized for producing suggestions. The aim is to explore and develop recommendation algorithms using past sales data. It covers the ideas behind content-based filtering, hybrid model recommendation, collaborative filtering and association rules for recommendation systems. For Product recommendation, three algorithms are used: the popularity product recommendation algorithm, the frequent pattern growth algorithm and Apriori algorithm. To increase sales and market response, the proposed recommendation model yields good recommendation outcomes. Additionally, it suggests particular products to potential clients. This paper describes the research environment for recommendation systems in grocery stores. From the observed recommendation models, Popular Product Recommendation does not require customer analysis. Among the models that compare the customer behavior, FP Growth serves better than Apriori model due to its faster execution time. This paper provided a very relevant and practical business transformation scenario that helps businesses in comparable circumstances change their business models.

Collaborative filtering and deep learning based recommendation system for cold start items

Recommendation systems have emerged as effective tools for mitigating information overload. Traditionally, recommendation systems employ various models such as Collaborative Filtering, Matrix Decomposition, and Logic Decomposition. Among these, Collaborative Filtering stands out due to its high efficiency. However, it encounters challenges related to cold start and sparse data. To address these challenges, the integration of Knowledge Graphs with recommendation systems has demonstrated significant advantages. This paper classifies Knowledge Graph-based recommendation systems into two categories: enhanced classical recommendation models and novel recommendation models integrated with Knowledge Graphs. We provide explanations for each category and compare them with traditional methods to draw conclusions. To inspire future research endeavors, this article identifies potential research areas and highlights unresolved issues.

Recommendationn system assist to provide trustworthy results to users. Hence, Recommendation systems are becoming very popular in Healthcare also. This article analyses an intelligence-based recommender system that uses classification or prediction methods to offer recommendations to HD patients in a telehealth environment. The system offers patient choice support regarding the need for medical testing based on analytics of each patient’s medical test results in records. Artificial intelligence technology-based systems might positively impact the human lifestyle of patients grief from CD (chronic diseases) by giving helpful HRSs (health recommendation systems). This article has analyzed hybridCNN and SVM models that give heart patients HD prediction and medical advice. The fuzzy in CNN is used to detect disease levels and helps generate recommendations. The initial section’s primary objective at developing a classification method that can verify the HD of a patient and predict it into different output classes like; silent ischemia, angina, chest pain, etc. The main motive of this research is to construct an AIbased recommendation system for HD patients. The simulation analysis for the research system is shown on a clinical database gathered through discussion with medical specialists from an identified hospital. The evaluation of the researched recommender model is calculated by accuracy, RMSE, SP, and SN as calculation events. The HRSs are created based on data gathered from a knowledge-based generated with the help of doctors. The research outcomes of the detection and recommender models are calculated and given accuracy with a Hybrid-CNN of 94.3 per cent.

Recently, deep learning models have been widely explored in recommender systems. Though having achieved remarkable success, the design of task-aware recommendation models usually requires manual feature engineering and architecture engineering from domain experts. To relieve those efforts, we explore the potential of neural architecture search (NAS) and introduce AMEIR for Automatic behavior Modeling, interaction Exploration and multi-layer perceptron (MLP) Investigation in the Recommender system. Specifically, AMEIR divides the complete recommendation models into three stages of behavior modeling, interaction exploration, MLP aggregation, and introduces a novel search space containing three tailored subspaces that cover most of the existing methods and thus allow for searching better models. To find the ideal architecture efficiently and effectively, AMEIR realizes the one-shot random search in recommendation progressively on the three stages and assembles the search results as the final outcome. The experiment over various scenarios reveals that AMEIR outperforms competitive baselines of elaborate manual design and leading algorithmic complex NAS methods with lower model complexity and comparable time cost, indicating efficacy, efficiency, and robustness of the proposed method.

Recommendation systems have been widely used in many internet communities, as they could provide users with personalized items based on the big data analysis technologies. Collaborative Filtering (CF) recommendation algorithm is a well-known and widely used technique in the recommendation systems. However, traditional collaborative filtering methods usually make predictions only based on the rating-based features, which could not make a better understanding of user preferences. Moreover, with the growing disclosure of online personal information, privacy protection is another important issue to be considered in dynamic cyber networks. Thus, a security-aware prediction mechanism is necessary to deal with the data privacy in recommendation systems. In view of these challenges, this paper proposes a privacy-aware multi-preference-based collaborative filtering recommendation method with Locality-Sensitive Hashing (LSH). Firstly, two types of user representations are obtained based on both rating features and latent factors. To protect the privacy data of users, both the LSH technique and differential privacy data perturbation mechanism are integrated into our recommendation model. Besides, the employment of LSH in establishing users' nearest neighbor matrix also improve the similarity searching efficiency of the CF recommendation algorithm. Final recommendation is based on a SGD-based hybrid collaborative prediction model. Experimental results show that our proposed algorithm can improve the accuracy of the recommendation system and reduce the time cost on the basis of ensuring user privacy.

A recommendation system can assist the user to compose an understanding of requirements and propose informed decisions from a lot of complicated knowledge. Recommendation from an analysis of sentiments seems to be a great challenge as user-generated content is represented using human language in several complicated ways. Many studies have focused on common fields such as reviews of electrical items, films, and restaurants, but not enough on health and medical issues. Sentiment analysis of healthcare in general and that of the drug experiences of individuals, in particular, may shed considerable light on how to focus on improving public health and reach the correct decision. In this paper, we design and implement a drug recommender system framework that applies sentiment analysis technologies on drug reviews. The objective of this research is to build a decision-making support platform to help patients to achieve more significant choices in drug selection. Firstly, we propose a sentimental measurement approach to drug reviews and generate ratings on drugs. Secondly, we take how much the drug reviews are useful to users, patient's conditions, and dictionary sentiment polarity of drug reviews into consideration. Then, we fuse those factors into the recommendation system to list appropriate medications. Experiments have been carried out using Decision Tree, K-Nearest Neighbors, and Linear Support Vector Classifier algorithm in rating generation and Hybrid model in recommendation based on the given open dataset. The analysis is carried out to tune the parameters for each algorithm in order to achieve greater performance. Finally, Linear Support Vector Classifier is selected for rating generation to obtain a good trade-off among model accuracy, model efficiency, and model scalability.

The Recommender System (RS) has obtained a pivotal role in e-commerce. To improve the performance of RS, review text information has been extensively utilized. However, it is still a challenge for RS to extract the most informative feature from a tremendous amount of reviews. Another significant issue is the modeling of user–item interaction, which is rarely considered to capture high- and low-order interactions simultaneously. In this paper, we design a multi-level attention mechanism to learn the usefulness of reviews and the significance of words by Deep Neural Networks (DNN). In addition, we develop a hybrid prediction structure that integrates Factorization Machine (FM) and DNN to model low-order user–item interactions as in FM and capture the high-order interactions as in DNN. Based on these two designs, we build a Multi-level Attentional and Hybrid-prediction-based Recommender (MAHR) model for recommendation. Extensive experiments on Amazon and Yelp datasets showed that our approach provides more accurate recommendations than the state-of-the-art recommendation approaches. Furthermore, the verification experiments and explainability study, including the visualization of attention modules and the review-usefulness prediction test, also validated the reasonability of our multi-level attention mechanism and hybrid prediction.