Joint Deep Model with Multi-Level Attention and Hybrid-Prediction for Recommendation.

Factorization Machines (FMs) are effective in incorporating side information to overcome the cold-start and data sparsity problems in recommender systems. Traditional FMs adopt the inner product to model the second-order interactions between different attributes, which are represented via feature vectors. The problem is that the inner product violates the triangle inequality property of feature vectors. As a result, it cannot well capture fine-grained attribute interactions, resulting in sub-optimal performance. Recently, the euclidean distance is exploited in FMs to replace the inner product and has delivered better performance. However, previous FM methods including the ones equipped with the euclidean distance all focus on the attribute-level interaction modeling, ignoring the critical intrinsic feature correlations inside attributes. Thereby, they fail to model the complex and rich interactions exhibited in the real-world data. To tackle this problem, in this paper, we propose a FM framework equipped with generalized metric learning techniques to better capture these feature correlations. In particular, based on this framework, we present a Mahalanobis distance and a deep neural network (DNN) methods, which can effectively model the linear and non-linear correlations between features, respectively. Besides, we design an efficient approach for simplifying the model functions. Experiments on several benchmark datasets demonstrate that our proposed framework outperforms several state-of-the-art baselines by a large margin. Moreover, we collect a new large-scale dataset on second-hand trading to justify the effectiveness of our method over cold-start and data sparsity problems in recommender systems.

In recent years, multidimensional data is becoming increasingly popular in recommender systems. For example, social tagging systems encourage users to employ user-defined keywords to help manage content in a personalized way. Recommender systems built upon social tagging systems utilize social tagging data to improve recommendation accuracy. Context-aware recommender systems incorporate context information (e.g. time, location, weather, etc.) into recommendation models to make recommendations more personalized. User profiles play a crucial role in recommender systems because user profiles provide the information about users’ information needs based on which personalized recommendations can be generated. Existing user profiling techniques for collaborative filtering (CF) recommender systems based on multidimensional data mostly analyze data through splitting the multidimensional relations into lower-dimensional relations. However, this leads to the loss of multidimensionality in user-item interactions. A different type of methods for addressing multidimensional data is factorization techniques such as tensor factorization (TF), which discover holistic latent relations between all dimensions in multidimensional data. However, TF models are not constructed to analyze any particular dimension of multidimensional data. But for many application domains, certain dimensions are of particular interest such as the user dimension in recommender systems. However, user profiles are not explicitly reflected in TF models. In this paper, we propose a unified approach to profile any particular dimension in multidimensional data based on partially-decomposed Tucker models (TM). We prove that the proposed profiling approach is a compatible higher-order extension of the matrix factorization based approaches. Moreover, we prove that the proposed TM-based profiles can be updated incrementally with new data without reconstructing the entire profiles. We integrate the proposed TF-based user/item profiling approach into neighborhood-based collaborative filtering recommenders for making top-N item recommendations. Extensive experiments have been conducted using real-world social tagging datasets to demonstrate that the CF recommenders integrated with the proposed user/item profiles outperform state-of-the-art CF recommendation approaches in terms of recommendation accuracy.

Recommender models are hard to evaluate, particularly under offline setting. In this article, we provide a comprehensive and critical analysis of the data leakage issue in recommender system offline evaluation. Data leakage is caused by not observing global timeline in evaluating recommenders e.g., train/test data split does not follow global timeline. As a result, a model learns from the user-item interactions that are not expected to be available at the prediction time. We first show the temporal dynamics of user-item interactions along global timeline, then explain why data leakage exists for collaborative filtering models. Through carefully designed experiments, we show that all models indeed recommend future items that are not available at the time point of a test instance, as the result of data leakage. The experiments are conducted with four widely used baseline models—BPR, NeuMF, SASRec, and LightGCN, on four popular offline datasets—MovieLens-25M, Yelp, Amazon-music, and Amazon-electronic, adopting leave-last-one-out data split. 1 We further show that data leakage does impact models’ recommendation accuracy. Their relative performance orders thus become unpredictable with different amount of leaked future data in training. To evaluate recommendation systems in a realistic manner in offline setting, we propose a timeline scheme, which calls for a revisit of the recommendation model design.

The cold-start problem in recommender systems has been facing a great challenge. Cross-domain recommendation can improve the performance of cold-start user recommendations in the target domain by using the rich information of users in the source domain. In cross-domain cold-start recommendation, users in target domain lack sufficient historical behaviors. Existing meta-learning-based methods depend on the feature distribution of training data and limit the adaptability in new tasks. To address these issues, we propose a meta-adversarial framework for cross-domain cold-start recommendation (MAFCDR) . Specifically, we employ a multi-level feature attention mechanism for independently learning the weights of long-term and short-term features to construct preferences of users in source domain. To migrate user representations, we train a meta-adversarial network that utilizes feature embeddings in the source domain as input and enhances the robustness and stability of the model. Then, the personalized bridge function transfers the user preferences in the source domain to the target domain. We build three cross-domain tasks using Amazon dataset and conduct extensive experiments, which demonstrate the effectiveness of the proposed model in cold-start user recommendation.

In an era of information explosion, recommender systems are vital tools to deliver personalized recommendations for users. The key of recommender systems is to forecast users' future behaviors based on previous user-item interactions. Due to their strong expressive power of capturing high-order connectivities in user-item interaction data, recent years have witnessed a rising interest in leveraging Graph Neural Networks (GNNs) to boost the prediction performance of recommender systems. Nonetheless, classic Matrix Factorization (MF) and Deep Neural Network (DNN) approaches still play an important role in real-world large-scale recommender systems due to their scalability advantages. Despite the existence of GNN-acceleration solutions, it remains an open question whether GNN-based recommender systems can scale as efficiently as classic MF and DNN methods. In this paper, we propose a Linear-Time Graph Neural Network (LTGNN) to scale up GNN-based recommender systems to achieve comparable scalability as classic MF approaches while maintaining GNNs' powerful expressiveness for superior prediction accuracy. Extensive experiments and ablation studies are presented to validate the effectiveness and scalability of the proposed algorithm. Our implementation based on PyTorch is available.

The building industry's high energy consumption and greenhouse gas emissions make accurate heating and cooling load prediction crucial.Traditional white box and black box methods have limitations.This paper explores Deep Neural Networks (DNNs) and the Deep -FM model, which combines Factorization Machines (FM) and DNNs.The DNN imitates the human brain, while Deep -FM explicitly captures high -order feature interactions and learns non -linear patterns.Using a dataset from the UCI machine learning repository with 768 data points and 10 variables related to building attributes, the performance of these two models was evaluated.Four metrics (MAE, MSE, MAPE, and R2) were used.The model structures were optimized, and the results show that Deep -FM is more effective for simpler models and limited data, while DNN excels in complex scenarios with sufficient data.The performance of both models improves with more training data.Deep -FM outperforms DNN in predicting heating and cooling loads under small training datasets.Additionally, deep learning algorithms show much higher accuracy than shallow learning algorithms in predicting cooling load.

English learning resources involve a large amount of language knowledge and semantic information, and users' English learning needs are multidimensional, and these needs will dynamically change with the learning process and time. It is prone to insufficient feedback and evaluation information, with a large amount of sparse information, making it difficult to accurately analyze user learning resource preferences. How to capture and adapt to the multi-dimensional demand characteristics of users in real-time is a challenge faced by recommendation systems. To this end, research is conducted on a sparse information filtering recommendation algorithm for English resource libraries that integrates multi-level interactive attention mechanisms. Filter sparse interval data through FCM membership threshold (0.2<μ<0.8), extract core English resource features, and then use a multi-level interactive attention mechanism to hierarchically extract preference features from the user layer (7 types of features such as age/interest) and resource layer (7 types of features such as listening/writing). After feature fusion, use Top-K method to calculate resource similarity and generate a recommendation list. Experiments have shown that on a dataset of 1500 resource items, 700 users, and 50000 ratings, the algorithm achieves significantly better performance in three key indicators: consistency in preference feature extraction (0.949-0.968), resource coverage (≥ 0.9), and conversion rate (96.83%) compared to the baseline model (LSTM model conversion rate 39.35%), and a 2.6-fold increase in detail page clicks. By dynamically capturing the user resource interaction relationship, the algorithm achieved an accurate proportion matching of 0.868 (with a deviation of 0.002) in the recommendation of listening resources, verifying its superiority in multi-dimensional dynamic demand scenarios.

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

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.

A dual learning-based recommendation approach

Gliomas, the most aggressive primary tumors in the central nervous system, are characterized by high morphological heterogeneity and diffusely infiltrating boundaries. Such complexity poses significant challenges for accurate segmentation in clinical practice. Although deep learning methods have shown promising results, they often struggle to achieve a satisfactory trade-off among precise boundary delineation, robust multi-scale feature representation, and computational efficiency, particularly when processing high-resolution three-dimensional (3D) magnetic resonance imaging (MRI) data. Therefore, the aim of this study is to develop a novel 3D segmentation framework that specifically addresses these challenges, thereby improving clinical utility in brain tumor analysis. To accomplish this, we propose a multi-level channel-spatial attention and light-weight scale-fusion network (MCSLF-Net), which integrates a multi-level channel-spatial attention mechanism (MCSAM) and a light-weight scale-fusion module. By strategically enhancing subtle boundary features while maintaining a compact network design, our approach seeks to achieve high accuracy in delineating complex glioma morphologies, reduce computational burden, and provide a more clinically feasible segmentation solution. We propose MCSLF-Net, a network integrating two key components: (I) MCSAM: by strategically inserting a 3D channel-spatial attention module at critical semantic layers, the network progressively emphasizes subtle, infiltrative edges and small, easily overlooked contours. This avoids reliance on an additional edge detection branch while enabling fine-grained localization in ambiguous transitional regions. (II) Light-weight scale fusion unit (LSFU): leveraging depth-wise separable convolutions combined with multi-scale atrous (dilated) convolutions, LSFU enhances computational efficiency and adapts to varying feature requirements at different network depths. In doing so, it effectively captures small infiltrative lesions as well as extensive tumor areas. By coupling these two modules, MCSLF-Net balances global contextual information with local fine-grained features, simultaneously reducing the computational burden typically associated with 3D medical image segmentation. Extensive experiments on the BraTS 2019, BraTS 2020, and BraTS 2021 datasets validated the effectiveness of our approach. On BraTS 2021, MCSLF-Net achieved a mean Dice similarity coefficient (DSC) of 0.8974 and a mean 95th percentile Hausdorff distance (HD95) of 2.52 mm. Notably, it excels in segmenting intricate transitional areas, including the enhancing tumor (ET) region and the tumor core (TC), thereby demonstrating superior boundary delineation and multi-scale feature fusion capabilities relative to existing methods. These findings underscore the clinical potential of deploying multi-level channel-spatial attention and light-weight multi-scale fusion strategies in high-precision 3D glioma segmentation. By striking an optimal balance among boundary accuracy, multi-scale feature capture, and computational efficiency, the proposed MCSLF-Net offers a practical framework for further advancements in automated brain tumor analysis and can be extended to a range of 3D medical image segmentation tasks.

Recommender system has received tremendous attention and has been studied by scholars in recent years due to its wide applications in different domains. With the in‐depth study and application of deep learning algorithms, deep neural network is gradually used in recommender systems. The success of modern recommender system mainly depends on the understanding and application of the context of recommendation requests. However, when leveraging deep learning algorithms for recommendation, the impact of context information such as recommendation time and location is often neglected. In this paper, a time‐aware convolutional neural network‐ (CNN‐) based personalized recommender system TC-PR is proposed. TC-PR actively recommends items that meet users’ interests by analyzing users’ features, items’ features, and users’ ratings, as well as users’ time context. Moreover, we use Tensorflow distributed open source framework to implement the proposed time‐aware CNN‐based recommendation algorithm which can effectively solve the problems of large data volume, large model, and slow speed of recommender system. The experimental results on the MovieLens‐1m real dataset show that the proposed TC-PR can effectively solve the cold‐start problem and greatly improve the speed of data processing and the accuracy of recommendation.

Recommender systems aim at predicting users’ future behaviors by learning the users’ personal information and historical behaviors. Unfortunately, training by the user’s raw data will inevitably cause the user’s private information to leak. Moreover, exiting privacy-preserving recommendation methods are based on matrix factorization with Local Differential Privacy (LDP). Because LDP performs poorly in small or multi-dimensional data sets, these methods do not work well with sparse or multi-dimensional features. Motivated by these questions, we propose a privacy-preserving recommendation model based on Deep neural networks and Factorization Machines (DeepFM), which can provide users with recommendations under the promise of protecting users’ privacy. In our recommendation model, each user randomly perturbs the calculated gradients to satisfy Condensed Local Differential Privacy (CLDP). The recommender system collects the perturbed gradients to train a recommendation model, which guarantees the safety of users’ private information and has excellent accuracy for the recommendation. Experiments on a real-world dataset show that the recommendation accuracy of our algorithm performs better than the existing methods.KeywordsPrivacy-preservingRecommender systemCondensed local differential privacyDeep neural networkFactorization machine

In today's rapidly evolving digital ecosystem, businesses are increasingly relying on personalized experiences to enhance customer satisfaction and ensure long-term success. One of the key drivers of this personalization is recommendation systems, which analyze vast amounts of customer data to suggest relevant products, services, and content. However, traditional recommendation methods such as collaborative filtering and content-based filtering have limitations in handling large-scale, dynamic datasets. This paper explores the integration of Artificial Intelligence (AI) through Deep Neural Networks (DNNs) to overcome these limitations, offering more accurate, scalable, and context-aware recommendations. DNNs have the capability to learn complex patterns and adapt to real-time changes in user behavior, making them ideal for enterprises dealing with diverse, multi-dimensional data. Despite the promising potential of DNN-based recommendation systems, challenges such as computational complexity, data privacy concerns, and model interpretability must be addressed. The paper provides a comprehensive framework for implementing DNN-driven recommendation systems in enterprise environments, addressing technical, operational, and ethical considerations. The findings of this study suggest that AI-driven systems, particularly those utilizing DNNs, can significantly improve customer engagement, optimize marketing strategies, and enhance conversion rates, ultimately leading to more personalized and effective customer interactions.

Recommendation is very crucial technique for social networking sites and business organizations. It provides suggestions based on users’ personalized interest and provide users with movies, books and topics links that would be most suitable for them. It can improve user effectiveness and business revenue by approximately 30%, if analyzed in intelligent manner. Social recommendation systems for traditional datasets are already analyzed by researchers and practitioners in detail. Several researchers have improved recommendation accuracy and throughput by using various innovative approaches. Deep learning has been proven to provide significant improvements in image processing and object recognition. It is machine learning technique where hidden layers are used to improve outcome. In traditional recommendation techniques, sparsity and cold start are limitations which are due to less user-item interactions. This can be removed by using deep learning models which can improve user-item matrix entries by using feature learning. In this paper, various models are explained with their applications. Readers can identify best suitable model from these deep learning models for recommendation based on their needs and incorporate in their techniques. When these recommendation systems are deployed on large scale of data, accuracy degrades significantly. Social big graph is most suitable for large scale social data. Further improvements for recommendations are explained with the use of large scale graph partitioning. MAE (Mean Absolute Error) and RMSE (Root Mean Squared Error) are used as evaluation parameters which are used to prove better recommendation accuracy. Epinions, MovieLens and FilmTrust datasets are also shown as most commonly used datasets for recommendation purpose.