Multimedia Recommender Systems Research Articles

Personalized Multimedia Recommendation Systems Using Higher-Order Tensor Singular-Value-Decomposition

Because of rapid growth of multimedia data over the Internet, the infobesity has been emerging in recent years. Many recommender systems (RSs) have been proposed using a variety of techniques, including artificial intelligence, machine learning, statistical analysis, and data mining. Nowadays, E-commerce companies such as Amazon, eBay, Netflix, etc. have applied RSs conveniently to provide better customer services than before by enabling customized multimedia-content recommendations. It is well known that classical RSs are almost all based on the user-item attribute matrices. However, this two-fold user-item relationship should be extended to the emerging many-fold relationship among an arbitrary number of attribute groups so that the corresponding new RSs may lead to better performances. Therefore, in this paper, we introduce a novel personalized multimedia recommender system (PMRS) according to our proposed new parallel higher-order tensor singular-value-decomposition (PHOTSVD) algorithm. This PHOTSVD algorithm is capable of accommodating an arbitrary number of entities such as social-media channels, users, queries, webpages, etc. in a robust RS. Furthermore, we also investigate the cold start and data bias problems critical to RSs. Finally, we compare the performances of our proposed new PMRS and three other existing tensor-based RSs over realworld data with respect to the well-adopted metric, namely normalized root-mean-square error (NRMSE). Our proposed new PMRS greatly outperforms the other three existing tensor-based RSs in terms of NRMSE.

Robust multimedia recommender system based on dynamic collaborative filtering and directed adversarial learning

Robust multimedia recommender system based on dynamic collaborative filtering and directed adversarial learning

Read between the interactions: Understanding non-interacted items for accurate multimedia recommendation

This paper addresses the problem of multimedia recommendation that additionally utilizes multimedia data, such as visual and textual modalities of items along with the user-item interaction information. Existing multimedia recommender systems assume that all the non-interacted items of a user have the same degree of negativity, thus regarding them as candidates for negative samples when training the model. However, this paper claims that a user?s non-interacted items do not have the same degree of negativity. We classify these non-interacted items of a user into two kinds of items with different characteristics: unknown and uninteresting items. Then, we propose a novel negative sampling technique that only considers the uninteresting items (i.e., rather than the unknown items) as candidates for negative samples. In addition, we show that using the multiple Bayesian personalized ranking (BPR) losses with both unknown and uninteresting items (i.e., all the noninteracted items) in existing multimedia recommendation methods is very effective in improving recommendation accuracy. By conducting extensive experiments with three real-world datasets, we show the superiority of our ideas. Our ideas can be easily and orthogonally applied to any multimedia recommender systems.

ALGORITHMIC SUPPORT OF THE WEB SERVICE RECOMMENDATION SYSTEM FOR LEARNING FOREIGN LANGUAGES

This work is devoted to the analysis of algorithmic support of multimedia content recommender systems and the development of a web service toincrease the efficiency of learning foreign languages using a recommender system that personalized the selection of educational content for the user.To form a list of necessary multimedia content, the main criteria of the recommender system were selected, the basic needs of users were identified,which the system should solve, since increasing the efficiency of learning a foreign language is achieved not only by choosing teaching methods, butalso by watching multimedia content, namely news, films, educational videos, clips, etc. Therefore, in order to form a list of the necessary multimediacontent, the main criteria of the recommender system were formed, the main needs of users were identified, which the system must solve. From theside of the method for implementing algorithmic support, various types of data filtering were considered, from modern technical methods to librariesto ensure the functionality of the system, and the algorithm based on hybrid filtering was chosen, in which known user ratings are used to predict thepreferences of another user. Functional requirements have been developed and a web service has been proposed that allows a comprehensive impact onuser learning when learning a foreign language, software implementation of which is made using Java Script, Python and additional libraries. Thisimplementation allows you to build a process for tracking changes in user requirements and transfer information to the database (DB) and, afteranalyzing the input data, change the proposed multimedia content to the user. In the course of further research, it is planned to conduct practicalexperiments, taking into account the specifics of certain methods of teaching foreign languages and the use of statistical data to assess the effectivenessof the algorithm of the proposed recommender system.

Emotion driven mood enhancing multimedia recommendation system using physiological signal

This paper exhibits a proposal for developing a multimedia recommendation system that is driven by the emotion of the user. The multimedia system is designed in such a way that the user will not have to manually browse for the audio/video to be played, it will be played automatically based on the emotion. The database comprises ECG signals collected from DECAF. Anger and sadness, emotions contributing to negative mood are considered. Discrete Wavelet Transform (DWT) was used for feature extraction and classification accuracy was determined using kNN, SVM. It was found that kNN classifies better than SVM. Realtime signals of one single healthy subject were collected based on a protocol built. Multimedia recommendation system operates automatically based on the classified emotion.

The multimedia recommendation algorithm based on probability graphical model

In the multimedia big data, the demand for personalized multimedia recommendation algorithm is increasing to ease the multimedia information overload. The multimedia recommendation system has been applied in various industries and has been playing a significant role. With the development of multimedia big data, developing multimedia recommendation algorithms can effectively be used in multimedia data. However, a large number of prevailing recommendation systems cannot meet the multimedia recommendation requirements, since they ignore the user-item interactions with multimedia content. This essay realizes the multimedia recommendation based on probability graphical model, to deal with the cold start and data sparsity involved in collaborative filtering recommendation, proposing that add the user tag to user-item model. The essay optimizes the multimedia recommendation algorithm based on undirected graphical model and tests it with singular value decomposition, clustering and Naive Bayes separately. The essay also builds the checklist recommendation model and experiments extensively for comparison with the conditional multimedia recommendation algorithm, by using PersonalRank algorithm based on random-walk to work out the weight coefficient of the user tag. At the same time, the essay enhances the probability-graph multimedia recommendation algorithm by dimensionality reduction and clustering, with the result of noticeably improved precision and recall.

Online incremental hierarchical classification resonance network

Hierarchical classification is imperative in that almost all objects are described in hierarchical semantics. If a classification method enables incremental class learning to learn new objects online, it will be practically used for real-time applications. In this sense, we propose online incremental hierarchical classification resonance network (OIHCRN) that enables online incremental class learning in hierarchical classification. OIHCRN has a structure that grows horizontally and vertically online according to object classes, so that a newly added object can be classified. By the proposed process of scale-preserving projection and prior label appending, OIHCRN reflects the class dependency between class levels and simultaneously normalizes the input vector online. Additionally, to reduce the model complexity and improve performance, two auxiliary strategies, named OIHCRN with class END and OIHCRN with differentiated class labels, are introduced. To demonstrate the effectiveness of OIHCRNs, experiments are carried out for benchmark datasets and then for a multimedia recommendation system.

Adversarial Training Towards Robust Multimedia Recommender System

With the prevalence of multimedia content on the Web, developing recommender solutions that can effectively leverage the rich signal in multimedia data is in urgent need. Owing to the success of deep neural networks in representation learning, recent advance on multimedia recommendation has largely focused on exploring deep learning methods to improve the recommendation accuracy. To date, however, there has been little effort to investigate the robustness of multimedia representation and its impact on the performance of multimedia recommendation. In this paper, we shed light on the robustness of multimedia recommender system. Using the state-of-the-art recommendation framework and deep image features, we demonstrate that the overall system is not robust, such that a small (but purposeful) perturbation on the input image will severely decrease the recommendation accuracy. This implies the possible weakness of multimedia recommender system in predicting user preference, and more importantly, the potential of improvement by enhancing its robustness. To this end, we propose a novel solution named Adversarial Multimedia Recommendation (AMR), which can lead to a more robust multimedia recommender model by using adversarial learning. The idea is to train the model to defend an adversary, which adds perturbations to the target image with the purpose of decreasing the model's accuracy. We conduct experiments on two representative multimedia recommendation tasks, namely, image recommendation and visually-aware product recommendation. Extensive results verify the positive effect of adversarial learning and demonstrate the effectiveness of our AMR method. Source codes are available in https://github.com/duxy-me/AMR.

A Resource Retrieval Method Of Multimedia Recommendation System Based On Deep Learning

In order to overcome the problem of too much time and low efficiency in retrieval caused by fuzzy representation of word vectors in multimedia recommendation system, this paper proposes a method of resource retrieval based on deep learning. Firstly, this method preprocesses multimedia resources and retrieval sentences, including morpheme segmentation, stop word processing, term normalisation, stem extraction, word vector processing. In the end, a deep retrieval model is built, to calculate the similarity between multimedia resources and retrieval statements, so as to achieve resource matching. The experimental results show that, while ensuring the accuracy of retrieval, the MAP value of short, medium and long retrieval statement fields is the maximum, which is 0.95, 0.92 and 0.88 respectively. The retrieval time is shortened by 0.3-0.6 s, which improves the retrieval efficiency.

A resource retrieval method of multimedia recommendation system based on deep learning

A resource retrieval method of multimedia recommendation system based on deep learning

A Generalized Evaluation Framework for Multimedia Recommender Systems

With the widespread availability of media technologies, such as real-time streaming, new Internet-of-Thing devices and smart phones, multimedia data are extensively increased and the big multimedia data rapidly spread over various social networks. This has created complexity and information overload for users to choose the suitable multimedia objects. Thus, different multimedia recommender systems have been emerging to help users find the useful multimedia objects that are possibly preferred by the user. However, the evaluation of these multimedia recommender systems is still in an ad-hoc stage. Given the distinct features of multimedia objects, the evaluation criteria adopted from the general recommender systems might not be effectively used to evaluate multimedia recommendations. In this paper, we therefore review and analyze the evaluation criteria that have been used in the previous multimedia recommender system papers. Based on the review, we propose a generalized evaluation framework to guide the researchers and practitioners to perform evaluations, especially user-centric evaluations, for multimedia recommender systems.

Incremental Class Learning for Hierarchical Classification

Objects can be described in hierarchical semantics, and people also perceive them this way. It leads to the need for hierarchical classification in machine learning. On the other hand, when a new data that belongs to a new class is given, the existing classification methods should be retrained for all data including the new data. To deal with these issues, we propose an adaptive resonance theory-supervised predictive mapping for hierarchical classification (ARTMAP-HC) network that allows incremental class learning for raw data without normalization in advance. Our proposed ARTMAP-HC is composed of hierarchically stacked modules, and each module incorporates two fuzzy ARTMAP networks. Regardless of the level of the class hierarchy and the number of classes for each level, ARTMAP-HC is able to incrementally learn sequentially added input data belonging to new classes. By using a novel online normalization process, ARTMAP-HC can classify the new data without prior knowledge of the maximum value of the dataset. By adopting the prior labels appending process, the class dependency between class hierarchy levels is reflected in ARTMAP-HC. The effectiveness of the proposed ARTMAP-HC is validated through experiments on hierarchical classification datasets. To demonstrate the applicability, ARTMAP-HC is applied to a multimedia recommendation system for digital storytelling.

A Multimedia Content Recommender System Using Table of Contents and Content-Based Filtering

In recent years, we have witnessed a rapid growth in the availability of digital multimedia on various application platforms and domains. Consequently, the problem of information overload has become more and more serious which causes a difficulty and time consuming for the users to locate and discovers the desired data. In this paper, we propose an architecture of a Multimedia Recommender system for Education videos that recommends relevant videos to the user based on Content-Based Filtering and the Table of Content of relevant books, that in turn will increase the dynamism and accuracy of the recommended multimedia data of their preference, preventing them from consuming time and effort in locating the desired video.

A Survey of Multimedia Recommender Systems: Challenges and Opportunities

Multimedia information has been extensively growing from a variety of sources such as cameras or video recorders. In order to select the useful multimedia objects, multimedia recommender system has been emerging as a tool to help users choose which multimedia objects might be interesting for them. However, given the complexity of multimedia objects, it is challenging to provide effective multimedia recommendations. In this paper, we therefore conduct a survey in both the multimedia information system and recommender system communities. We further focus on the works that span the two communities, especially the research on multimedia recommender systems. Based on our review, we propose a set of research challenges, which can be used to implicate the future research directions for multimedia recommender systems. For each research challenge, we have also provided the insights of how to perform the follow-up research.

CHIS: A big data infrastructure to manage digital cultural items

In this paper, we describe CHIS (Cultural Heritage Information System), a big data infrastructure that can be used to query, browse, analyze and process digital contents related to cultural heritage from a set of heterogeneous and distributed repositories. CHIS is characterized by the following technical features: capability to gather information from distributed and heterogeneous data sources (e.g., Sensor Networks, Social Media Networks, Digital Libraries and Archives, Multimedia Collections, Web Data Services, etc.); advanced data management techniques and technologies; ability to provide useful and personalized data to users based on their preferences and context; advanced information retrieval facilities, data analytics and other utilities/services, according to the SOA paradigm. By means of a set of ad-hoc APIs, and value-added data processing and analytics services, our system can support several applications: mobile multimedia guides for cultural environments, web portals to promote the cultural heritage of a given organization, multimedia recommender and storytelling systems and so on. We discuss the main ideas that characterize the system, showing its use for several applications.

Personalized Multimedia Recommendations for Cloud-Integrated Cyber-Physical Systems

Portable smart devices have paved the way for accessing and capturing different types of multimedia contents with human interactions, leading to the emergence of cyber-physical systems (CPSs). Although the massive data collected from these physical terminals can contribute to the improvement of their quality of lives by building smart communities, CPSs intensify the information overload problem. Therefore, plenty of research efforts have been paid to develop multimedia recommender systems. However, most existing research activities neglect its time-varying features due to system dynamics, i.e., not only the amount of input data constantly grows, but also the change of user behaviors and system operating environment. In order to sustain the high accuracy of recommendations, the system in a CPS has to be updated regularly. However, the more often the update proceeds, the more the cost of other computational resources. To this end, in this paper, we propose an adaptive recommender system by using feedback control frameworks in CPSs. The proposed solution continuously monitors its changes and estimates the loss of performance (in terms of accuracy) to overcome the data aging problem and justify if the current revisiting ratio between the new and old items can still accurately reflect current user behavior. Theoretical analysis and extensive results by using a real data set in a cloud setting are supplemented to show the advantages of the proposed system.

Multimedia recommendation and transmission system based on cloud platform

This paper presents a movie recommendation system according to scores that the users provide. In view of the movie evaluation system, the impacts of access control and multimedia security are analyzed, and a secure hybrid cloud storage architecture is presented. Mobile-Edge Computing (MEC) technology is used in the public cloud which guarantees the high efficiency requirements of the transmission of the multimedia content. The processes of the system including registration, user login, role assignment, data encryption and data decryption are also described. At last, the performance of the proposed scheme is analyzed which further shows that the various possible attacks can be mitigated via the proposed system.

IJARCCE - Computer and Communication Engineering

Trust is an important part in a social network from security point of view. Online video sharing systems is the most popular and provide features that allow users to post a video in a web page. These features provide opportunities for a user to introduce polluted content into the system. Spammersmay post an unrelated video aiming at increasing the likelihood of the responsebeing viewed by a larger number of users. Multimedia recommendation system recommends video based on the user behavior which reduces network overhead and speed up the recommendation process. The proposed approach can recommend desired services with high precision, high recall and low response delay. To avoid the explosion of networkoverhead, user-behavior-based clustering is performed. If unrelated content is displayed in our web page, then we spam that video content. If more users spam all video content from same provider, then the provider will be deleted from server.

Cloud-Based Mobile Multimedia Recommendation System With User Behavior Information

Facing massive multimedia services and contents in the Internet, mobile users usually waste a lot of time to obtain their interests. Therefore, various context-aware recommendation systems have been proposed. Most of those proposed systems deploy a large number of context collectors at terminals and access networks. However, the context collecting and exchanging result in heavy network overhead, and the context processing consumes huge computation. In this paper, a cloud-based mobile multimedia recommendation system which can reduce network overhead and speed up the recommendation process is proposed. The users are classified into several groups according to their context types and values. With the accurate classification rules, the context details are not necessary to compute, and the huge network overhead is reduced. Moreover, user contexts, user relationships, and user profiles are collected from video-sharing websites to generate multimedia recommendation rules based on the Hadoop platform. When a new user request arrives, the rules will be extended and optimized to make real-time recommendation. The results show that the proposed approach can recommend desired services with high precision, high recall, and low response delay.

Social and Content Hybrid Image Recommender System for Mobile Social Networks

One of the advantages of social networks is the possibility to socialize and personalize the content created or shared by the users. In mobile social networks, where the devices have limited capabilities in terms of screen size and computing power, Multimedia Recommender Systems help to present the most relevant content to the users, depending on their tastes, relationships and profile. Previous recommender systems are not able to cope with the uncertainty of automated tagging and are knowledge domain dependant. In addition, the instantiation of a recommender in this domain should cope with problems arising from the collaborative filtering inherent nature (cold start, banana problem, large number of users to run, etc.). The solution presented in this paper addresses the abovementioned problems by proposing a hybrid image recommender system, which combines collaborative filtering (social techniques) with content-based techniques, leaving the user the liberty to give these processes a personal weight. It takes into account aesthetics and the formal characteristics of the images to overcome the problems of current techniques, improving the performance of existing systems to create a mobile social networks recommender with a high degree of adaptation to any kind of user.