Heterogeneous Graph Embedding for Cross-Domain Recommendation Through Adversarial Learning

Real-world complex networks are inherently heterogeneous; they have different types of nodes, attributes, and relationships. In recent years, various methods have been proposed to automatically learn how to encode the structural and semantic information contained in heterogeneous information networks (HINs) into low-dimensional embeddings; this task is called heterogeneous network embedding (HNE). Efficient HNE techniques can benefit various HIN-based machine learning tasks such as node classification, recommender systems, and information retrieval. Here, we provide a comprehensive survey of key advancements in the area of HNE. First, we define an encoder-decoder-based HNE model taxonomy. Then, we systematically overview, compare, and summarize various state-of-the-art HNE models and analyze the advantages and disadvantages of various model categories to identify more potentially competitive HNE frameworks. We also summarize the application fields, benchmark datasets, open source tools, and performance evaluation in the HNE area. Finally, we discuss open issues and suggest promising future directions. We anticipate that this survey will provide deep insights into research in the field of HNE.

Interpretable answer retrieval based on heterogeneous network embedding

Recommendation System has become more customized than ever. However, it suffers from cold start and data sparsity issues. Cross Domain Recommendation (CDR) has been proposed as a promising and feasible approach which yields recommendations in the destination domain based on the characteristic and personalized preferences of the users in the source domain. Since the fondness varies for each user, their taste bridge would also be different. Most of the well-known existing models make use of common preference bridge when it comes to transferring preferences for all the users in the system. Personalized Preference Transfer (PPT) for personalizing user preferences using Characteristic Encoder (CE) where user characteristic is mapped to initializing maps is pro-posed for cold start users and from books for warm start users. In this model, a Meta net-work with user attributes is used to attain the bridge preferences for individual users. The Deep Neural Network (DNN) which is a neural network model is used to provide optimized recommendations to the users. The accuracy and efficiency of personalized user preferences transfer is determined based on the comparisons from different models and train/test ratios, making use of two evaluation metrics MAE (Mean Absolute Error) and RMSD (Root Mean Squared Deviation).

Sliced Wasserstein based Canonical Correlation Analysis for Cross-Domain Recommendation

In recent years, the recommender system has been widely used in online platforms, which can extract useful information from giant volumes of data and recommend suitable items to the user according to user preferences. However, the recommender system usually suffers from sparsity and cold-start problems. Cross-domain recommendation, as a particular example of transfer learning, has been used to solve the aforementioned problems. However, many existing cross-domain recommendation approaches are based on matrix factorization, which can only learn the shallow and linear characteristics of users and items. Therefore, in this article, we propose a novel autoencoder framework with an attention mechanism (AAM) for cross-domain recommendation, which can transfer and fuse information between different domains and make a more accurate rating prediction. The main idea of the proposed framework lies in utilizing autoencoder, multilayer perceptron, and self-attention to extract user and item features, learn the user and item-latent factors, and fuse the user-latent factors from different domains, respectively. In addition, to learn the affinity of the user-latent factors between different domains in a multiaspect level, we also strengthen the self-attention mechanism by using multihead self-attention and propose AAM++. Experiments conducted on two real-world datasets empirically demonstrate that our proposed methods outperform the state-of-the-art methods in cross-domain recommendation and AAM++ performs better than AAM on sparse and large-scale datasets.

Recommendation system can predict the ratings of users to items by leveraging machine learning algorithms. The use of recommendation systems is common in e-commerce websites now-a-days. Since enormous amounts of data including users’ click streams, purchase history, demographics, social networking comments and user-item ratings are stored in e-commerce systems databases, the volume of the data is getting bigger at high speed, and the data is sparse. However, the recommendations and predictions must be made in real time, enabling to bring enormous benefits to human beings. Apache spark is well suited for applications which require high speed query of data, transformation and analytics results. Therefore, the recommendation system developed in this research is implemented on Apache Spark. Also, the matrix factorization using Alternating Least Squares (ALS) algorithm which is a type of collaborative filtering is used to solve overfitting issues in sparse data and increases prediction accuracy. The overfitting problem arises in the data as the user-item rating matrix is sparse. In this research a recommendation system for e-commerce using alternating least squares (ALS) matrix factorization method on Apache Spark MLlib is developed. The research shows that the RMSE value is significantly reduced using ALS matrix factorization method and the RMSE is 0.870. Consequently, it is shown that the ALS algorithm is suitable for training explicit feedback data set where users provide ratings for items.

The conventional single-target Cross-Domain Recommendation (CDR) only improves the recommendation accuracy on a target domain with the help of a source domain (with relatively richer information). In contrast, the novel dual-target CDR has been proposed to improve the recommendation accuracies on both domains simultaneously. However, dual-target CDR faces two new challenges: (1) how to generate more representative user and item embeddings, and (2) how to effectively optimize the user/item embeddings on each domain. To address these challenges, in this paper, we propose a graphical and attentional framework, called GA-DTCDR. In GA-DTCDR, we first construct two separate heterogeneous graphs based on the rating and content information from two domains to generate more representative user and item embeddings. Then, we propose an element-wise attention mechanism to effectively combine the embeddings of common users learned from both domains. Both steps significantly enhance the quality of user and item embeddings and thus improve the recommendation accuracy on each domain. Extensive experiments conducted on four real-world datasets demonstrate that GA-DTCDR significantly outperforms the state-of-the-art approaches.

The world is filled with information and getting the right information is a challenging task for internet users and online buyers. Recommender system helps internet users to get their information in a short span of time. It acts as an information extraction system that works behind users to perform their search easier. The recommender system comes under user’s content or item based search, similar users browsing behavior called collaborative and combination of both known as a hybrid. Here collaborative-based approach is adopted which recommends items to their users based on their past browsing behavior. In this article, the User-Item-Rating matrix is formulated concerning user personal profile, rating of the product, and reviews given by the users during their previous browsing history. In this research, user location is considered as an important attribute to group similar users. It also attempts to suppress the scalability and sparsity problems of the traditional collaborative filtering approach. So, the User-Item-Rating (UIR) matrix has considered the location, ratings and reviews for future recommendation. The Navie Bayes classifier algorithm is used to provide accurate topmost recommendations to internet users. The data set is taken from the MovieLens and IMDb database. The accuracy of the recommender system is measured based on the main metric f-measure. The experimental result has proven the improvement of the recommender system with the mentioned added attributes.

The tagging system provides users with a platform to express their preferences as they annotate terms or keywords to items. Tag information is a bridge between two domains for transferring knowledge and helping to alleviate the data sparsity problem, which is a crucial and challenging problem in most recommender systems. Existing methods incorporate correlations extracted from overlapping tags at a lexical level in cross-domain recommendation, but they neglect semantical relationships between different tags, which impairs prediction accuracy in the target domain. To solve this challenging problem, we propose a cross-domain recommendation method with semantic correlation in tagging systems. This method automatically captures the semantic relationships between non-identical tags and applies them to the recommendation. The word2vec technique is used to learn the latent representations of tags. Semantically equivalent tags are then grouped to form a joint embedding space comprised of tag clusters. This embedding space serves as the bridge between domains. By mapping users and items from both the source and target domains into the same embedding space, similar users or items across domains can be identified. Thus, the recommendation in a sparse target domain is improved by transferring knowledge through correlated users and items. Experimental results with three datasets on six cross-domain recommendation tasks demonstrate that the proposed method exploits the semantic links from tags in two domains and outperforms five benchmarks in prediction accuracy. The results indicate that transferring knowledge through tags semantics is feasible and effective.

Cold start problem is one of the most challenging and long-standing problems in recommender systems, and cross-domain recommendation (CDR) methods are effective for tackling it. Most cold-start related CDR methods require training a mapping function between high-dimensional embedding space using overlapping user data. However, the overlapping data is scarce in many recommendation tasks, which makes it difficult to train the mapping function. In this paper, we propose a new approach for CDR, which aims to alleviate the training difficulty. The proposed method can be viewed as a special parameterization of the mapping function without hurting expressiveness, which makes use of non-overlapping user data and leads to effective optimization. Extensive experiments on two real-world CDR tasks are performed to evaluate the proposed method. In the case that there are few overlapping data, the proposed method outperforms the existed state-of-the-art method by 14% (relative improvement).

Web systems that provide the same functionality usually share a certain amount of items. This makes it possible to combine data from different websites to improve recommendation quality, known as the cross-domain recommendation task. Despite many research efforts on this task, the main drawback is that they largely assume the data of different systems can be fully shared. Such an assumption is unrealistic - different systems are typically operated by different companies, and it may violate business privacy policy to directly share user behavior data since it is highly sensitive. In this work, we consider a more practical scenario to perform cross-domain recommendation. To avoid the leak of user privacy during the data sharing process, we consider sharing only the information of the item side, rather than user behavior data. Specifically, we transfer the item embeddings across domains, making it easier for two companies to reach a consensus (e.g., legal policy) on data sharing since the data to be shared is user-irrelevant and has no explicit semantics. To distill useful signals from transferred item embeddings, we rely on the strong representation power of neural networks and develop a new method named as NATR (short for Neural Attentive Transfer Recommendation). We perform extensive experiments on two real-world datasets, demonstrating that NATR achieves similar or even better performance than traditional cross-domain recommendation methods that directly share user-relevant data. Further insights are provided on the efficacy of NATR in using the transferred item embeddings to alleviate the data sparsity issue.

This paper proposes a cross-domain consumer recommendation algorithm model based on the combination of heterogeneous information network embedding and neural networks. The model first builds a heterogeneous information network and embeds various types of nodes and their complex relationships, such as users, products, and fields, into a low-dimensional vector space to fully capture the potential semantic associations of cross-domain data. Next, a feature alignment module is designed to alleviate the impact of data feature differences in different fields by normalizing and mapping embedded features, thereby ensuring consistent expression of cross-domain features. Finally, the model uses multi-layer perceptron (MLP) or graph neural network (GNN) to perform non-linear fusion and modeling of multi-modal features, and introduces an attention mechanism to dynamically capture the weight of key features to further improve user cross-domain preferences predictive ability. In order to verify the effectiveness of the model, this article carried out comparative experiments in multiple cross-domain recommendation scenarios. The results show that the model proposed in this article is significantly better than the traditional single-mode model in core indicators such as recommendation accuracy, recall rate and F1 value, modal recommendation algorithm and simple multi-modal fusion method.

With the increasing problem of information overload, recommenders are essential on many websites to help users find interesting content quickly. User feedback is the most valuable form of model feedback, and the form and how user feedback is used affects the accuracy, coverage and experience of recommendations. Thus, this paper reviews feedback types in recommender systems: explicit, implicit, and deliberate implicit feedback, and discusses their pros and cons in terms of accuracy, ease-of-use, and user effort. Additionally, it discusses challenges recommendation models face when using feedback, like data sparsity, bias reinforcement, and fairness limitations. Meanwhile, it proposes strategies to address these challenges by integrating multi-source feedback, fairness-aware modeling, and interpretability enhancement to boost user confidence. The results show gains in accuracy, coverage, and fairness. In addition, it provides exploratory analyses on dynamic feedback and cross-domain recommendations, offering insights for modeling and applying multi-type feedback in future recommender systems.

Phenotype prediction aims to infer the traits of living organisms based on genomics data, which has important applications in biology such as cancer subtype diagnosis and crop breedings. Traditional phenotype prediction approaches only learn the low-dimensional representation of samples, but ignore the interaction between biomolecules and cannot use the topology structure of heterogeneous molecular networks. Furthermore, most of them lack interpretability and do not effectively identify key biomolecules associated with phenotypes. In this paper, we propose a heterogeneous network embedding based solution (PhenoHNE) to predict phenotype by fusing topology information of molecules. PhenoHNE firstly utilizes variational graph autoencode (VGAE) to obtain the embedding representation of molecules in the heterogeneous network. Secondly, PhenoHNE adopts multilayer perceptron (MLP) with attention mechanism to learn sample representation. Finally, it fuses molecular embedding representation and sample representation to predict the phenotype of samples. In this way, the genetics information of heterogeneous molecular network can further guide the feature learning of samples and improve the prediction performance. Experimental results on human and maize datasets confirm that PhenoHNE outperforms competitive methods by a large margin under different evaluation protocols, and it also can effectively identify the key molecules associated with phenotypes of interests.

Making accurate recommendations for cold-start users has been a longstanding and critical challenge for recommender systems (RS). Cross-domain recommendations (CDR) offer a solution to tackle such a cold-start problem when there is no sufficient data for the users who have rarely used the system. An effective approach in CDR is to leverage the knowledge (e.g., user representations) learned from a related but different domain and transfer it to the target domain. Fine-tuning works as an effective transfer learning technique for this objective, which adapts the parameters of a pre-trained model from the source domain to the target domain. However, current methods are mainly based on the global fine-tuning strategy: the decision of which layers of the pre-trained model to freeze or fine-tune is taken for all users in the target domain. In this paper, we argue that users in RS are personalized and should have their own fine-tuning policies for better preference transfer learning. As such, we propose a novel User-specific Adaptive Fine-tuning method (UAF), selecting which layers of the pre-trained network to fine-tune, on a per-user basis. Specifically, we devise a policy network with three alternative strategies to automatically decide which layers to be fine-tuned and which layers to have their parameters frozen for each user. Extensive experiments show that the proposed UAF exhibits significantly better and more robust performance for user cold-start recommendation.