Search

Most existing RGB-D salient object detection (SOD) methods tend to achieve higher performance by integrating additional modules, such as feature enhancement and edge generation. There is no doubt that these modules will inevitably produce feature redundancy and performance degradation. To this end, we exquisitely design a cross-modal fusion and progressive decoding network (termed CPNet) to achieve RGB-D SOD tasks. The designed network structure only includes three indispensable parts: feature encoding, feature fusion and feature decoding. Specifically, in the feature encoding part, we adopt a two-stream Swin Transformer encoder to extract multi-level and multi-scale features from RGB images and depth images respectively to model global information. In the feature fusion part, we design a cross-modal attention fusion module, which can leverage the attention mechanism to fuse multi-modality and multi-level features. In the feature decoding part, we design a progressive decoder to gradually fuse low-level features and filter noise information to accurately predict salient objects. Extensive experimental results on 6 benchmarks demonstrated that our network surpasses 12 state-of-the-art methods in terms of four metrics. In addition, it is also verified that for the RGB-D SOD task, the addition of the feature enhancement module and the edge generation module is not conducive to improving the detection performance under this framework, which provides new insights into the salient object detection task. Our codes are available at https://github.com/hu-xh/CPNet. [ABSTRACT FROM AUTHOR]

The primary goal of recommender systems is to identify and propose items that users might find appealing. A large number of these systems are heavily dependent on explicit interactions between the user and the item, which can often be infrequent. In this work, we introduce a unique model known as Context-Aware Metric Learning (CAML), designed to enhance the effectiveness of recommendations. The CAML model utilizes an attentive autoencoder to extract latent features from contextual context and incorporates these features into a metric learning framework. In particular, these extracted features act as a Gaussian prior for the embeddings of the items, thereby enhancing the precision of their positioning in the latent space. This integration not only boosts the precision of the recommendations but also increases computational efficiency, rendering CAML appropriate for both offline and online application scenarios. Our model's evaluation on two real-world datasets reveals that it outperforms several existing baseline models, including those that do not incorporate contextual information such as CML and CPE, as well as other contextual recommendation models like CDL, CATA, and CML+F. [ABSTRACT FROM AUTHOR]

To recommend useful information to users more efficiently, this paper proposes a dual-path recommendation algorithm which combines multilayer Convolutional Neural Network (CNN) and attention-enhanced long short-term memory network (Attention-LSTM). Firstly, the matrix factorisation technique is used for learning the long-term preferences of users. Secondly, a dual-path network based on CNN and LSTM is constructed to perform feature extraction on the rating matrix. The dual-path network can learn the long-term preferences of users while capturing their dynamic preferences in changing preferences. The algorithm is tested on the public dataset MovieLens-1M, and the MAE value reflects the accuracy of the algorithm. [ABSTRACT FROM AUTHOR]

Sequential recommendation systems in cold-start scenarios aim to provide recommendations as accurately as possible for users with sparse behavior, which is a challenging issue in this field. Recently, meta-learning algorithms have been introduced into the cold-start recommendations and have obtained some better results. However, most of these meta-learning-based methods require auxiliary information or knowledge from other fields, and do not model the third-order relationship between users and their sequential interactive items, which could result in models lacking the ability to capture dynamic transitions of user preferences. In addition, the traditional meta-learning-based methods ignore the diversity of user preferences, which limits the performance improvement in sequential recommendation systems in the cold-start scenarios. To address the above problems, a new sequential recommendation model is proposed for cold-start scenarios which combines meta-learning and attention mechanism. In the proposed model, a third-order interaction modeling paradigm of "Item-User-Item" (IUI) is proposed firstly to obtain dynamic transitions of user preferences in the item space. Then, an embedding strategy is used to embed the user behavior sequences and transition information of user preferences from the training task into a spatial self-attentive (SS) recommendation model. Moreover, a meta-learning-based parameter training approach is presented, where a task processor (TP) is designed to improve the universality of parameter initialization. At last, some experiments are conducted on three real-world datasets, and the evaluation metrics HR@10 and NDCG@10 of the proposed model improve by about 5.1% and 5.6% compared with the baseline, respectively. The experimental results show that the proposed model in this paper achieves the best recommendation results in comparison with existing models. [ABSTRACT FROM AUTHOR]

This paper addresses the challenges of data sparsity and personalization limitations inherent in current recommendation systems when processing extensive academic paper datasets. To overcome these issues, the present work introduces an innovative recommendation model that integrates the wealth of structured information from knowledge graphs and refines the amalgamation of temporal and relational data. By applying attention mechanisms and neural network technologies, the model thoroughly explores the text characteristics of papers and the evolving patterns of user behaviors. Additionally, the model elevates the accuracy and personalization of recommendations by meticulously examining citation patterns among papers and the networks of author collaboration. The experimental findings show that the present model surpasses baseline models on all evaluation metrics, thereby enhancing the precision and personalization of academic paper recommendations. [ABSTRACT FROM AUTHOR]

The relationship between users and items, which cannot be recovered by traditional techniques, can be extracted by the recommendation algorithm based on the graph convolution network. The current simple linear combination of these algorithms may not be sufficient to extract the complex structure of user interaction data. This paper presents a new approach to address such issues, utilizing the graph convolution network to extract association relations. The proposed approach mainly includes three modules: Embedding layer, forward propagation layer, and score prediction layer. The embedding layer models users and items according to their interaction information and generates initial feature vectors as input for the forward propagation layer. The forward propagation layer designs two parallel graph convolution networks with self-connections, which extract higher-order association relevance from users and items separately by multi-layer graph convolution. Furthermore, the forward propagation layer integrates the attention factor to assign different weights among the hop neighbors of the graph convolution network fusion, capturing more comprehensive association relevance between users and items as input for the score prediction layer. The score prediction layer introduces MLP (multi-layer perceptron) to conduct nonlinear feature interaction between users and items, respectively. Finally, the prediction score of users to items is obtained. The recall rate and normalized discounted cumulative gain were used as evaluation indexes. The proposed approach effectively integrates higher-order information in user entries, and experimental analysis demonstrates its superiority over the existing algorithms. [ABSTRACT FROM AUTHOR]

In recent years, graph neural networks (GNNs) have been widely applied in recommender systems. However, existing recommendation algorithms based on GNNs still face challenges in node aggregation and feature extraction processes because they often lack the ability to capture the interactions between users and items, as well as users' multiple intentions. This hinders accurate understanding of users' needs. To address the aforementioned issues, we propose a recommendation model called multi-intent-aware interactive graph convolutional network (Multi-IAIGCN). This model is capable of integrating multiple user intents and adopts an interactive convolution approach to better capture the information on the interaction between users and items. First, before the interaction between users and items begins, user intents are divided and mapped into a graph. Next, interactive convolutions are applied to the user and item trees. Finally, by aggregating different features of user intents, predictions of user preferences are made. Extensive experiments on three publicly available datasets demonstrate that Multi-IAIGCN outperforms existing state-of-the-art methods or can achieve results comparable to those of existing state-of-the-art methods in terms of recall and NDCG, thus verifying the effectiveness of Multi-IAIGCN. [ABSTRACT FROM AUTHOR]

The primary goal of recommender systems is to identify and propose items that users might find appealing. A large number of these systems are heavily dependent on explicit interactions between the user and the item, which can often be infrequent. In this work, we introduce a unique model known as Context-Aware Metric Learning (CAML), designed to enhance the effectiveness of recommendations. The CAML model utilizes an attentive autoencoder to extract latent features from contextual context and incorporates these features into a metric learning framework. In particular, these extracted features act as a Gaussian prior for the embeddings of the items, thereby enhancing the precision of their positioning in the latent space. This integration not only boosts the precision of the recommendations but also increases computational efficiency, rendering CAML appropriate for both offline and online application scenarios. Our model’s evaluation on two real-world datasets reveals that it outperforms several existing baseline models, including those that do not incorporate contextual information such as CML and CPE, as well as other contextual recommendation models like CDL, CATA, and CML+F.

Recommendation systems are an effective solution to deal with information overload, particularly in the e-commerce sector, in which sequential recommendation is extensively utilized. Sequential recommendations aim to acquire users' interests and provide accurate recommendations by analyzing users' historical interaction sequences. To improve recommendation performance, it is vital to take into account the long- and short-term interests of users. Despite significant advancements in this domain, some issues need to be addressed. Conventional sequential recommendation models typically express each item with a uniform embedding, ignoring evolutionary patterns among item attributes, such as category, brand, and price. Moreover, these models often model users' long- and short-term interests independently, failing to adequately address the issues of interest drift and short-term interest evolution. This study proposes a new model, the Feature-aware Long-Short Interest Evolution Network (FLSIE), to address the above-mentioned issues. Specifically, the model uses explicit feature embedding to represent item attribute information and employs a two-dimensional (2D) attention mechanism to distinguish the significance of individual features in a specific item and the relevance of each item in the interaction sequence. Furthermore, to avoid the issue of interest drift, the model employs a long-term interest guidance mechanism to enhance the representation of short-term interest and adopts a gated recurrent unit with attentional update gate to model the dynamic evolution of users' short-term interest. Experimental results indicate that our presented model outperforms existing methods on three real-world datasets. [ABSTRACT FROM AUTHOR]

Knowledge graphs (KGs) play a vital role in enhancing search results and recommendation systems. With the rapid increase in the size of KGs, they are becoming inaccurate and incomplete. This problem can be solved by the KG completion methods, of which graph attention network (GAT)-based methods stand out because of their superior performance. However, existing GAT-based KG completion methods often suffer from overfitting issues when dealing with heterogeneous KGs, primarily due to the unbalanced number of samples. Additionally, these methods demonstrate poor performance in predicting the tail (head) entity that shares the same relation and head (tail) entity with others. To solve these problems, we propose GATH, a novel GAT-based method designed for Heterogeneous KGs. GATH incorporates two separate attention network modules that work synergistically to predict the missing entities. We also introduce novel encoding and feature transformation approaches, enabling the robust performance of GATH in scenarios with imbalanced samples. Comprehensive experiments are conducted to evaluate GATH's performance. Compared with the existing state-of-the-art GAT-based model on Hits@10 and MRR metrics, our model improves performance by 5.2% and 5.2% on the FB15K-237 dataset and by 4.5% and 14.6% on the WN18RR dataset, respectively. [ABSTRACT FROM AUTHOR]

Occlusion or posture change of the face in natural scenes has typical asymmetry; however, an asymmetric face plays a key part in the lack of information available for facial expression recognition. To solve the problem of low accuracy of asymmetric facial expression recognition, this paper proposes a fusion of channel global features and a spatial local information expression recognition network called the "Channel–Spatial Integration Network" (CSINet). First, to extract the underlying detail information and deepen the network, the attention residual module with a redundant information filtering function is designed, and the backbone feature-extraction network is constituted by module stacking. Second, considering the loss of information in the local key area of face occlusion, the channel–spatial fusion structure is constructed, and the channel features and spatial features are combined to enhance the accuracy of occluded facial recognition. Finally, before the full connection layer, more local spatial information is embedded into the global channel information to capture the relationship between different channel–spatial targets, which improves the accuracy of feature expression. Experimental results on the natural scene facial expression data sets RAF-DB and FERPlus show that the recognition accuracies of the modeling approach proposed in this paper are 89.67% and 90.83%, which are 13.24% and 11.52% higher than that of the baseline network ResNet50, respectively. Compared with the latest facial expression recognition methods such as CVT, PACVT, etc., the method in this paper obtains better evaluation results of masked facial expression recognition, which provides certain theoretical and technical references for daily facial emotion analysis and human–computer interaction applications. [ABSTRACT FROM AUTHOR]

Recently, most studies in the field have focused on integrating reviews behind ratings to improve recommendation performance. However, two main problems remain (1) Most works use a unified data form and the same processing method to address the user and the item reviews, regardless of their essential differences. (2) Most works only adopt simple concatenation operation when constructing user-item interaction, thus ignoring the multilevel relationship between the user and the item, which may lead to suboptimal recommendation performance. In this paper, we propose a novel Asymmetric Multi-Level Interactive Attention Network (AMLIAN) integrating reviews for item recommendation. AMLIAN can predict precise ratings to help the user make better and faster decisions. Specifically, to address the essential difference between the user and the item reviews, AMLIAN uses the asymmetric network to construct user and item features using different data forms (document-level and review-level). To learn more personalized user-item interaction, the user ID and item ID and some processed features of user reviews and item reviews are respectively used for multilevel relationships. Experiments on five real-world datasets show that AMLIAN significantly outperforms state-of-the-art methods. [ABSTRACT FROM AUTHOR]

Recommender systems are widely used in e-commerce, movies, music, social media, and other fields because of their personalized recommendation functions. The recommendation algorithm is used to capture user preferences, item characteristics, and the items that users are interested in are recommended to users. Matrix factorization is widely used in collaborative filtering algorithms because of its simplicity and efficiency. However, the simple dot-product method cannot establish a nonlinear relationship between user latent features and item latent features or make full use of their personalized information. The model of a neural network combined with an attention mechanism can effectively establish a nonlinear relationship between the potential features of users and items and improve the recommendation accuracy of the model. However, it is difficult for the general attention mechanism algorithm to solve the problem of attention interaction when the number of features between the users and items is not the same. To solve the above problems, this paper proposes an attention interaction matrix factorization (AIMF) model. The AIMF model adopts a symmetric structure using MLP calculation. This structure can simultaneously extract the nonlinear features of user latent features and item latent features, thus reducing the computation time of the model. In addition, an improved attention algorithm named slide-attention is included in the model. The algorithm uses the sliding query method to obtain the user's attention to the latent features of the item and solves the interaction problem among different dimensions of the user, and the latent features of the item. [ABSTRACT FROM AUTHOR]

In order to alleviate the inherent information differences between different user social spaces and interest spaces in recommendation systems and the problem of ignoring high-order neighbors, this paper proposed AGCRSR. Firstly, the algorithm used a mapping matrix in the embedding layer to convert the initial feature vectors into adaptive embedding. Secondly, it intro- duced an attention mechanism to aggregate different aspects of user embeddings, and used a graph convolutional network to li- nearly learn the potential representations of users and items. Finally, it used the adaptive module to aggregate user representations and predict the final recommendation results of users for the project using the inner product function. This paper conducted comparative experiments on the LastFM and Ciao datasets and compared with other baseline algorithms. The experimental results show that the recommendation performance of the AGCRSR is significantly improved compared to other algorithms. [ABSTRACT FROM AUTHOR]

Heterogeneous information networks (HINs) contain a rich network structure and semantic information, which makes them commonly used in recommendation systems. However, most of the existing HIN-based recommendation systems rely on meta-paths for information extraction, lack meta-path information supplements, and rarely learn complex structure information in heterogeneous graphs. To address these issues, we develop a novel recommendation algorithm that integrates the attention mechanism, meta-paths, and neighbor node information (AMNRec). In the heterogeneous information network, the missing information of the meta-path is supplemented by extracting the information of users and items' neighbor nodes. The rich interactions between nodes are captured through convolution, and the embedded representation of nodes and meta-paths is obtained through the attention mechanism. TOP-N recommendation is completed by combining users, items, neighbor nodes, and meta-paths. Experiments on three public datasets show that AMNRec not only has the best recommendation performance but also has good interpretability of the recommendation results compared with the six recommendation benchmark algorithms. [ABSTRACT FROM AUTHOR]

The expansion of available information in location-based social networks (LBSNs) has led to information overload, making it urgent to discover users' next point-of-interest (POI). Some existing works only consider certain modal information in LBSNs and do not transform them into high-dimensional structures, which hinders the alleviation of the data sparsity problem. Moreover, many approaches rely solely on social relationships, making it difficult to recommend POIs to new users without association information. To tackle these challenges, we propose a social- and spatial–temporal-aware next point-of-Interest (SSTP) recommendation model. SSTP uses two feature encoders based on self-attention mechanism and gate recurrent unit to model users' check-in enhancement sequence hierarchically. We also design a random neighborhood sampling approach to mine user social relationships, thus alleviating the user cold start problem. Finally, we propose a geographical-aware graph attention network to learn the sensitivity of users to distance. Extensive experiments on two real-world datasets show that SSTP outperforms state-of-the-art models, improving Hit@k by 2.26–6.55 % and MAP@k by 3.49–6.55 % . Moreover, SSTP has better performance on sparse data, with an average improvement of 6.09 % on the Hit@k. The code can be downloaded at https://github.com/Rih0/sstp. [ABSTRACT FROM AUTHOR]

Many researchers use AI to improve the accuracy of early diagnostic techniques. However, as a result of the tumor's uneven shape, fuzzy borders and too few data, existing tumor segmentation methods do not propose accurate segmentation results. We innovative introduces the prior knowledge learned to filter the noise information and guide the final network to generate a more accurate segmentation model. First, we introduce a classification network with an attention block to highlight the potential location of the brain tumor and also obtain the rough diagnosis result as the prior knowledge. Second, we provide a novel image fusion network consisting of a transformer with cross attention to merge tumor localization information with brain MRI images. Third, we propose a novel multilayer transformer experience information fusion network to combine the classic U‐Net network to handle the guiding of prior knowledge. The higher performance of the suggested method is demonstrated by comparison with contemporary methods. [ABSTRACT FROM AUTHOR]

Graph neural networks (GNNs) aim to learn well-trained representations in a lower-dimension space for downstream tasks while preserving the topological structures. In recent years, attention mechanism, which is brilliant in the fields of natural language processing and computer vision, is introduced to GNNs to adaptively select the discriminative features and automatically filter the noisy information. To the best of our knowledge, due to the fast-paced advances in this domain, a systematic overview of attention-based GNNs is still missing. To fill this gap, this paper aims to provide a comprehensive survey on recent advances in attention-based GNNs. Firstly, we propose a novel two-level taxonomy for attention-based GNNs from the perspective of development history and architectural perspectives. Specifically, the upper level reveals the three developmental stages of attention-based GNNs, including graph recurrent attention networks, graph attention networks, and graph transformers. The lower level focuses on various typical architectures of each stage. Secondly, we review these attention-based methods following the proposed taxonomy in detail and summarize the advantages and disadvantages of various models. A model characteristics table is also provided for a more comprehensive comparison. Thirdly, we share our thoughts on some open issues and future directions of attention-based GNNs. We hope this survey will provide researchers with an up-to-date reference regarding applications of attention-based GNNs. In addition, to cope with the rapid development in this field, we intend to share the relevant latest papers as an open resource at https://github.com/sunxiaobei/awesome-attention-based-gnns. [ABSTRACT FROM AUTHOR]

Microblogging services, such as Twitter and Sina Weibo, are popular media for information dissemination in cyberspace. Predicting this information dissemination (i.e., users' retweet behavior) is essential in various applications, such as applications for content recommendations and public opinion analyses. Most state-of-the-art models use deep representation methods to learn about the features of the input variables (e.g., social relationships and retweet history) and obtain predictions based on the association between the features and retweet behavior. However, these models ignore the inherent confounding association, which can cause models to learn spurious relationships in the data, thereby decreasing their generalizability. To solve this problem, this paper devises a novel causal-enriched deep attention (CEDA) framework for (1) evaluating the causal effects of input variables on retweet behaviors by blocking the confounding association and (2) assigning appropriate attention weights to different variables with causal effects when training models and obtaining predictions. Specifically, the CEDA uses a hidden variable to represent factors that are difficult to observe but can influence user behavior. This ensures that the model can more realistically portray a given microblogging service. Moreover, the score function of the CEDA can overcome the over-suppression problem that may arise from the causal inference. It avoids the loss of predictive information due to the elimination of the confounding association. The results of testing the CEDA on several public datasets indicate that the CEDA improves the performance (F1-score) of a retweet prediction model by an average of 3.73%. [ABSTRACT FROM AUTHOR]

Context-aware recommender systems are intended primarily to consider the circumstances under which a user encounters an item to provide better-personalized recommendations. Users acquire point-of-interest, movies, products, and various online resources as suggestions. Classical collaborative filtering algorithms are shown to be satisfactory in a variety of recommendation activities processes, but cannot often capture complicated interactions between item and user, along with sparsity and cold start constraints. Hence it becomes a surge to apply a deep learning-based recommender model owing to its dynamic modeling potential and sustained success in other fields of application. In this work, a trust-based attentive contextual denoising autoencoder (TACDA) for enhanced Top-N context-aware recommendation is proposed. Specifically, the TCADA model takes the sparse preference of the user that is integrated with trust data as input into the autoencoder to prevail over the cold start and sparsity obstacle and efficiently accumulates the context condition into the model via attention framework. Thereby, the attention technique is used to encode context features into a latent space of the user's trust data that is integrated with their preferences, which interconnects personalized context circumstances with the active user's choice to deliver recommendations suited to that active user. Experiments conducted on Epinions, Caio, and LibraryThing datasets make it obvious the efficiency of the TACDA model persistently outperforms the state-of-the-art methods. [ABSTRACT FROM AUTHOR]

The recommendation system helps users select satisfactory products and services to make reasonable decisions. In recent years, most methods have introduced social information into the recommendation system to improve recommendation accuracy. Most social recommendations only consider that users are affected by historical items and social information. But users are affected by multiple potential factors that cannot be explicitly distinguished when making decisions, such as consumption experience and life cycle related to users, social consumption culture, and social roles related to social friends, and they are mixed with each other. To describe the above scenario, we propose a model that Utilizes the influence of multiple potential factors for Social Recommendation (UimSRec). Specifically, we simulate the influence of different potential factors on users in the form of latent semantics, associate the potential influence with user representation, and use the user latent representation to model preferences. In addition, this paper uses the attention mechanism to adaptively assign weights to multiple influencing factors and common influencing factors. The experimental results on three real datasets show that modeling various potential impact information and their relationships can significantly improve the recommendation performance. The code is available at (https://github.com/qinkaili/UimSRec). [ABSTRACT FROM AUTHOR]

Sequential recommendation based on amulti-interest framework aims to analyze different aspects of interest based on historical interactions and generate predictions of a user's potential interest in a list of items. Most existing methods only focus on what are themultiple interests behind interactions but neglect the evolution of user interests over time. To explore the impact of temporal dynamics on interest extraction, this paper explicitly models the timestamp with amulti-interest network and proposes a time-highlighted network to learn user preferences, which considers not only the interests at different moments but also the possible trends of interest over time. More specifically, the time intervals between historical interactions and prediction moments are first mapped to vectors. Meanwhile, a time-attentive aggregation layer is designed to capture the trends of items in the sequence over time, where the time intervals are seen as additional information to distinguish the importance of different neighbors. Then, the learned items' transition trends are aggregated with the items themselves by a gated unit. Finally, a self-attention network is deployed to capture multiple interests with the obtained temporal information vectors. Extensive experiments are carried out based on three real-world datasets and the results convincingly establish the superiority of the proposed method over other state-of-the-art baselines in terms of model performance. [ABSTRACT FROM AUTHOR]

Hashtags are the keywords that describe the theme of social media content and have become very popular in influence marketing and trending topics. In recent years, hashtag prediction has become a hot topic in AI research to help users with automatic hashtag recommendations by capturing the theme of the post. Most of the previous work mainly focused only on textual information, but many microblog posts contain not only text but also the corresponding images. This work explores both image-text features of the microblog post. Inspired by the self-attention mechanism of the transformer in natural language processing, the visual-linguistics pre-train model with transfer learning also outperforms many downstream tasks that require image and text inputs. However, most of the existing models for multimodal hashtag recommendation are based on the traditional co-attention mechanism. This paper investigates the cross-modality transformer LXMERT for multimodal hashtag prediction for developing LXMERT4Hashtag, a cross-modality representation learning transformer model for hashtag prediction. It is a large-scale transformer model that consists of three encoders: a language encoder, an object encoder, and a cross-modality encoder. We evaluate the presented approach on dataset InstaNY100K. Experimental results show that our model is competitive and achieves impressive results, including precision of 50.5% vs 46.12%, recall of 44.02% vs 38.93%, and F1-score of 47.04% vs 42.22% compared to the existing state-of-the-art baseline model. [ABSTRACT FROM AUTHOR]

Predicting the popularity of online news is essential for news providers and recommendation systems. Time series, content and meta-feature are important features in news popularity prediction. However, there is a lack of exploration of how to integrate them effectively into a deep learning model and how effective and valuable they are to the model’s performance. This work proposes a novel deep learning model named Multiple Features Dynamic Fusion (MFDF) for news popularity prediction. For modeling time series, long short-term memory networks and attention-based convolution neural networks are used to capture long-term trends and short-term fluctuations of online news popularity. The typical convolution neural network gets headline semantic representation for modeling news headlines. In addition, a hierarchical attention network is exploited to extract news content semantic representation while using the latent Dirichlet allocation model to get the subject distribution of news as a semantic supplement. A factorization machine is employed to model the interaction relationship between metafeatures. Considering the role of these features at different stages, the proposed model exploits a time-based attention fusion layer to fuse multiple features dynamically. During the training phase, this work designs a loss function based on Newton’s cooling law to train the model better. Extensive experiments on the real-world dataset from Toutiao confirm the effectiveness of the dynamic fusion of multiple features and demonstrate significant performance improvements over state-of-the-art news prediction techniques. [ABSTRACT FROM AUTHOR]

Recently, object detection, which is focused on images with normal illumination levels, has achieved great success. However, the accuracy of object detection is reduced in suboptimal environments due to the images plagued by noise and low contrast. For boosting the performance of object-detection tasks under low-illumination conditions, we propose three modules for improvement: (1) the low-level feature attention (LFA) module learns to focus on the regional feature information of the object in the low-illumination environment, highlighting important features and filtering noisy information; (2) the feature fusion neck (FFN) obtains enriched feature information by fusing the feature information of the feature map after backbone; (3) the context-spatial decoupling head (CSDH) enables the classification head to focus on contextual semantic information so that the localization head obtains richer spatial details. Extensive experiments show that our algorithm realizing end-to-end detection shows good performance in low-illumination images. [ABSTRACT FROM AUTHOR]

In recent years, MKR has attracted increasing attention due to its ability to enhance the accuracy of recommendation systems through cooperation between the RS tasks and the KGE tasks, allowing for complementarity of the information. However, there are still three challenging issues: historical behavior preferences, missing data, and knowledge graph completion. To tackle these challenging problems, we propose MDAR, a multi-task learning approach that combines DeepFM with an attention mechanism (DeepAFM) and a relation-fused multi-head graph attention network (RMGAT). Firstly, we propose to leverage the attention mechanism in the DeepAFM to distinguish the importance of different features for target prediction by assigning different weights to different interaction features of the user and the item, which solves the first problem. Secondly, we introduce deep neural networks (DNNs) to extract the deep semantic information in the cross-compressed units by obtaining the high-dimensional features of the interactions between the RS task and the KG task to solve the second problem. Lastly, we design a multi-head graph attention network for relationship fusion (RMGAT) in the KGE task, which learns entity representations through the different contributions of the neighbors by aggregating the relationships into the attention network of the knowledge graph and by obtaining information about the neighbors with different importance for different relationships, effectively solving the third problem. Through experimenting on real-world public datasets, we demonstrate that MDAR obtained substantial results over state-of-the-art baselines for recommendations for movie, book, and music datasets. Our results underscore the effectiveness of MDAR and its potential to advance recommendation systems in various domains. [ABSTRACT FROM AUTHOR]

Point-of-Interest (POI) recommender systems can effectively assist users to find their preferred POIs. Recent studies mainly focus on extracting users' dynamic context from their check-in behaviors and using attention mechanism to capture different influence of context information for predicting their real-time requirements. However, the existing methods mainly focus on learning the weights of different POI as well as their correlations in the check-in sequences. In addition, these methods still suffer from limited performance, especially when the interaction data are sparse. In this paper, we propose a C ontext- and C ategory-aware D ouble S elf-A ttention (CCDSA) model for POI recommendation to explore and capture users' contextual preferences in two different aspects collaboratively, including the fine-grained preference for POI in check-in behaviors and the coarse-grained preference for category. Specifically, we first design a double self-attention mechanism module to learn the users' preferences for both POI and category in specific context. Then we combine users' check-in behaviors with POIs' category information to alleviate data sparsity problem in context-aware recommendation. Finally, we leverage context and category information to perform personalized POI recommendation. In particular, we devise an improved version of CCDSA, i.e., CCDSA+, which further replaces the self-attention mechanism with the sparse self-attention mechanism for improving training efficiency. The experimental results on four real-world datasets, Foursquare-NY, Foursquare-TKY, Weeplaces-NY and Weeplaces-SF show that the proposed models, CCDSA and CCDSA+, outperform the baselines. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Chongqing University of Posts & Telecommunications (Natural Science Edition) is the property of Chongqing University of Posts & Telecommunications and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

To solve the information overload issue and enhance the user experience of various web applications, recommender systems aim to better model user interests and preferences. Knowledge Graphs (KGs), consisting of real-world objective facts and fruitful entities, play a vital role in recommender systems. Recently, a technological trend has been to develop end-to-end Graph Neural Networks (GNNs)-based knowledge-aware recommendation (a.k.a., Knowledge Graph Recommendation, KGR) models. Unfortunately, current GNNs-based KGR approaches focus on how to capture high-order feature information on KGs while neglecting the following two crucial limitations: 1) The explicitly high-order feature interaction and fusion mechanism and 2) The valid user intent modelling mechanism. As such, these issues lead to insufficient user/item representation learning capability and unsatisfactory KGR performance. In this work, we present a novel Knowledge-enhanced Re commendation with F eature I nteraction and Intent-aware Attention Networks (FIRE) to address the latent intent modelling and high-order feature interaction deficiencies ignored by existing KGR methods. Based on the prototype user/item representation learning leveraging the GNNs-based approach, our model offers the following major improvements: One is the innovative use of Convolutional Neural Networks (CNNs) that perform vertical convolutional (a.k.a., bit-level convolutional) and horizontal convolutional (a.k.a., vector-level convolutional) processes to model multi-granular high-order feature interactions to enhance item-side representation learning. Another is to model users' latent intent factors by utilizing a two-level attention mechanism (i.e., node- and intent-level attention mechanism) to enhance user-side representation learning. Extensive experiments on three KGs domain public datasets demonstrate that our method outperforms the existing state-of-the-art baseline. Last but not least, numerous ablation- and model studies demystify the working mechanism and elucidate the plausibility of the proposed model. [ABSTRACT FROM AUTHOR]

In the process of image shooting, due to the influence of angle, distance, complex scenes, illumination intensity, and other factors, small targets and occluded targets will inevitably appear in the image. These targets have few effective pixels, few features, and no obvious features, which makes it difficult to extract their effective features and easily leads to false detection, missed detection, and repeated detection, thus affecting the performance of target detection models. To solve this problem, an improved faster region convolutional neural network (RCNN) algorithm integrating the convolutional block attention module (CBAM) and feature pyramid network (FPN) (CF-RCNN) is proposed to improve the detection and recognition accuracy of small-sized, occluded, or truncated objects in complex scenes. Firstly, it incorporates the CBAM attention mechanism in the feature extraction network in combination with the information filtered by spatial and channel attention modules, focusing on local efficient information of the feature image, which improves the detection ability in the face of obscured or truncated objects. Secondly, it introduces the FPN feature pyramid structure, and links high-level and bottom-level feature data to obtain high-resolution and strong semantic data to enhance the detection effect for small-sized objects. Finally, it optimizes non-maximum suppression (NMS) to compensate for the shortcomings of conventional NMS that mistakenly eliminates overlapping detection frames. The experimental results show that the mean average precision (MAP) of target detection of the improved algorithm on PASCAL VOC2012 public datasets is improved to 76.2%, which is 13.9 percentage points higher than those of the commonly used Faster RCNN and other algorithms. It is better than the commonly used small-sample target detection algorithm. [ABSTRACT FROM AUTHOR]

The emergence of recommender system is aimed at solving the problems brought by information explosion to human life and even the development of human society. As a traditional recommendation technique, collaborative filtering often encounters sparsity and cold start problems in many recommendation scenarios. Therefore, researchers have found that the introduction of side information can solve these problems to a certain extent and improve the performance of recommender systems. The knowledge graph is a heterogeneous graph that contains rich semantic relationships among items. The Multi-Perspective Learning based on Transformer Knowledge Graph Enhanced Recommendation (MPL-TransKR) proposed in this paper uses the knowledge graph as the side information for input and introduces the multi-head self-attention mechanism by reasonably combining the transformer idea. While learning the high-order neighborhood information of the items, the long-distance information between the items is captured, and the weight value is assigned to the user through the attention mechanism to strengthen the user representation to realize user-item multi-perspective learning to enhance the performance of the recommendation model. Through extensive experiments using public datasets, we demonstrated that MPL-TransKR performs well in book and music recommendations, surpassing state-of-the-art baselines on several metrics.

Copyright of Journal of Computer Engineering & Applications is the property of Beijing Journal of Computer Engineering & Applications Journal Co Ltd. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

The capture of user interest hidden in the user behavior sequence is a hot research direction of recommendation algorithms in recent years. The traditional sequence prediction model uses the last product clicked by the user as the target, and establishes the association between user behavior and the target product, but does not fully dig out the sequence relationship between user sequences. This study improves on the traditional DIN model, uses continuous behavior over a period of time as the target vector, uses the transformer structure to complete the sequence-to-sequence prediction task, and further extracts and utilizes the user's deep interest in the user behavior sequence, and it is recommended in conjunction with DIN as an auxiliary feature. The experimental results on the Amazon book and the electronic data sets show that the DIN-based hybrid recommendation model proposed in this study increases the AUC index of the original DIN model by about 0.7% and 1.9%, respectively. It can be seen that the hybrid recommendation based on user behavior sequence prediction can play a certain auxiliary role in the multi-feature recommendation system. In addition, the influence of user sequence length on the model results is also explored. [ABSTRACT FROM AUTHOR]

Knowledge graphs as external information has become one of the mainstream directions of current recommendation systems. Various knowledge-graph-representation methods have been proposed to promote the development of knowledge graphs in related fields. Knowledge-graph-embedding methods can learn entity information and complex relationships between the entities in knowledge graphs. Furthermore, recently proposed graph neural networks can learn higher-order representations of entities and relationships in knowledge graphs. Therefore, the complete presentation in the knowledge graph enriches the item information and alleviates the cold start of the recommendation process and too-sparse data. However, the knowledge graph's entire entity and relation representation in personalized recommendation tasks will introduce unnecessary noise information for different users. To learn the entity-relationship presentation in the knowledge graph while effectively removing noise information, we innovatively propose a model named k n o w l e d g e — e n h a n c e d   h i e r a r c h i c a l   g r a p h   c a p s u l e   n e t w o r k (KHGCN), which can extract node embeddings in graphs while learning the hierarchical structure of graphs. Our model eliminates noisy entities and relationship representations in the knowledge graph by the entity disentangling for the recommendation and introduces the attentive mechanism to strengthen the knowledge-graph aggregation. Our model learns the presentation of entity relationships by an original graph capsule network. The capsule neural networks represent the structured information between the entities more completely. We validate the proposed model on real-world datasets, and the validation results demonstrate the model's effectiveness. [ABSTRACT FROM AUTHOR]

In recent years, there has been a growing interest in developing next point-of-interest (POI) recommendation systems in both industry and academia. However, current POI recommendation strategies suffer from the lack of sufficient mixing of details of the features related to individual users and their corresponding contexts. To overcome this issue, we propose a deep learning model based on an attention mechanism in this study. The suggested technique employs an attention mechanism that focuses on the pattern's friendship, which is responsible for concentrating on the relevant features related to individual users. To compute context-aware similarities among diverse users, our model employs six features of each user as inputs, including user ID, hour, month, day, minute, and second of visiting time, which explore the influences of both spatial and temporal features for the users. In addition, we incorporate geographical information into our attention mechanism by creating an eccentricity score. Specifically, we map the trajectory of each user to a shape, such as a circle, triangle, or rectangle, each of which has a different eccentricity value. This attention-based mechanism is evaluated on two widely used datasets, and experimental outcomes prove a noteworthy improvement of our model over the state-of-the-art strategies for POI recommendation. [ABSTRACT FROM AUTHOR]

With the rapid development of Internet technology, people can learn all kinds of travel information anytime and anywhere. However, the serious information overload causes travelers to be unable to make accurate and reasonable travel routes that meet tourists' tastes for a while, thus reducing the quality of travel. The recommendation system as the mainstream solution to the information explosion of two means has received the attention of the majority of scholars and industry. Based on the research theory of tourist route recommendation, this paper analyzes the characteristics of attractions, factors affecting travelers' travel experience when touring attractions and factors affecting travelers' travel experience along tourist routes. Furthermore, we propose a tourist route recommendation model that meets tourists' preferences. Then, this paper uses the graph neural network (GNN) algorithm to build a framework for tourist route recommendations based on the GNN using the relationship of preference and commonality existing among groups, tourists and attractions. The GNN algorithm is optimized and improved using multiple graphs and an attention mechanism. Finally, the effectiveness of this paper's algorithm is verified by conducting experiments on different data sets. [ABSTRACT FROM AUTHOR]

In the field of knowledge-aware recommendation, embedding propagation methods can mine the structured information of the knowledge graph and also can obtained the semantic information between n-hop entities. However, with the increase of the spread, the embedding propagation method will lose part of the structure information, the user representation obtained in this way is coarse-grained and cannot adequately characterize the user's preferences. Aiming at these problems, this paper proposed an user preference network of fusing knowledge-aware and time-aware (FKTUPN). Firstly, FKTUPN generated the first-order embedding of the knowledge graph in the embedding-layer, and separately modeled the fine-grained user prefe-rence. Then it divided the time slot by random walk and performed temporal context encoding to get contextual information of KG. Subsequently, the contribution of the representation was distinguished and propagated through the knowledge-aware attention mechanism. Finally, it aggregated users and items of different propagation levels in the prediction layer, and obtained the recommendation result after multiplying these two representations. Experiments show that result on three datasets: Amazon-book, Last-FM and Yelp is superior to the baseline model in terms of recall@20 and NDCG@20, fine-grained user preference representation and full utilization of structured information can effectively improve the accuracy of recommendations. [ABSTRACT FROM AUTHOR]

In modern days, making recommendation for news articles poses a great challenge due to vast amount of online information. However, providing personalized recommendations from news articles, which are the sources of condense textual information is not a trivial task. A recommendation system needs to understand both the textual information of a news article, and the user contexts in terms of long-term and temporary preferences via the user's historic records. Unfortunately, many existing methods do not possess the capability to meet such need. In this work, we propose a neural deep news recommendation model called CupMar, that not only is able to learn the user-profile representation in different contexts, but also is able to leverage the multi-aspects properties of a news article to provide accurate, personalized news recommendations to users. The main components of our CupMar approach include the News Encoder and the User-Profile Encoder. Specifically, the News Encoder uses multiple properties such as news category, knowledge entity, title and body content with advanced neural network layers to derive informative news representation, while the User-Profile Encoder looks through a user's browsed news, infers both of her long-term and recent preference contexts to encode a user representation, and finds the most relevant candidate news for her. We evaluate our CupMar model with extensive experiments on the popular Microsoft News Dataset (MIND), and demonstrate the strong performance of our approach. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Henan Agricultural Sciences is the property of Editorial Board of Journal of Henan Agricultural Sciences and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

The existing knowledge graph embedding (KGE) method has achieved good performance in recommendation systems. However, the relevancy degree among entities reduces gradually along the spread in the knowledge graph. Focusing on the explicit and implicit relationships among entities, this paper proposes an attention knowledge network combining explicit and implicit information (AKNEI) to effectively capture and exactly describe the correlation between entities in the knowledge graph. First, we design an information-sharing layer (ISL) to realize information sharing between projects and entities through implicit interaction. We innovatively propose a cross-feature fusion module to extract high-order feature information in the model. At the same time, this paper uses the attention mechanism to solve the problem of the decline of information relevance in the process of knowledge graph propagation. Finally, the features of KGE and cross feature fusion module are integrated into the end-to-end learning framework, the item information in the recommendation task and the knowledge graph entity information are interacted implicitly and explicitly, and the characteristics between them are automatically learned. We performed extensive experiments on multiple public datasets that include movies, music, and books. According to the experimental results, our model has a great improvement in performance compared with the latest baseline. [ABSTRACT FROM AUTHOR]

Recommendation system utilizes user-item interactions and information on user's attributes to infer the user's interests and use them to make recommendations for the user. Graph neural network(GNN) has become more widely used in recommendation systems in recent years, because of their ability to naturally integrate node information and topology. However, most of the current recommendation methods based on graph structure only focus on a single recommendation domain (for example,session-based recommendation models or social recommendation models), without taking into account both the user's behavior information and the user's social relationships; Furthermore, session-based recommendation models usually use a recurrent neural network (RNN) to model user session, while RNN only focuses on the short-term impact of sessions and cannot cover all the information of sessions. Therefore, this paper proposes a novel session-based social recommendation model called GNNRec, which first utilizes gated graph neural network (GGNN) to represent users' session information, and then uses graph attention network (GAT) to aggregate social information of users and friends on social networks to effectively model users' interests. In this paper, experiments are conducted on two large datasets——Douban and Epinions, and the results show that the GNNRec model performs significantly better than current mainstream recommendation models. [ABSTRACT FROM AUTHOR]

Bearing fault diagnosis for equipment-safe operation has a crucial role. In recent years, more achievements have been made in bearing fault diagnosis. However, for the fault diagnosis model, the representation and sensitivity of bearing fault features have a great influence on the diagnosis output results; thus, the attention mechanism is particularly important for the selection of features. However, global attention focuses on all sequences, which is computationally expensive and not ideal for fault diagnosis tasks. The local attention mechanism ignores the relationship between non-adjacent sequences. To address the respective shortcomings of global attention and local attention, an adaptive sparse attention network is proposed in this paper to filter fault-sensitive information by soft threshold filtering. In addition, the effects of different signal representation domains on fault diagnosis results are investigated to filter out signal representation forms with better performance. Finally, the proposed adaptive sparse attention network is applied to cross-working conditions diagnosis of bearings. The adaptive sparse attention mechanism focuses on the signal characteristics of different frequency bands for different fault types. The proposed network model achieves better overall performance when comparing the cross-conditions diagnosis accuracy and model convergence speed. [ABSTRACT FROM AUTHOR]

To address the problem of poor recommendation accuracy in the session recommendation model, this paper proposed a capsule-based graph convolutional disentanglement for session-aware recommendation (CGCD) method.Specifically, it used disentangled learning technique to transition item embedding into factor embedding based on multiple sub-channels. Then, it used a capsule dynamic fusion strategy to aggregate different factors to obtain a new item embedding.In addition, it used a multi-head attention mechanism to assign weights to each item in the session. Finally, it aggregated the item embedding with other items in the current session according to the assigned weights, and then generated an accurate session representation to achieve item recommendation. Experiments on two publicly available real datasets show that the proposed model can improve the Pre@10, Pre@20, MRR@10 and MRR@20 by 5.17%, 2.99%, 6.56% and 2.94% on average of the recommended performance, which verifies the effectiveness and efficiency of the method in this paper. [ABSTRACT FROM AUTHOR]

When modeling user-item interaction sequences to extract sequential patterns, current recommender systems face the dual issues of a) long-distance dependencies in conjunction with b) high levels of noise. In addition, with the complexity of current recommendation model architectures there is a significant increase in computation time. Therefore, these models cannot meet the requirement of fast response needed in application scenarios such as online advertising. To deal with these issues, we propose a Knowledge Distilled Attention-based Latent Information Extraction Network for Sequential user behavior (KD-ALIENS). In this model structure, user and item attributes and history are utilized to model the latent information from high-order feature interactions in conjunction with user sequential historical behavior. With regard to the issues of long-distance dependency and noise, we have adopted the self-attention mechanism to learn the sequential patterns between items in a user-item interaction history. With regard to the issue of a complex model architecture which cannot meet the requirement of fast response, the use of model compression and acceleration is realized by: (a) use of a knowledge-distilled teacher and student module, wherein the complex teacher module extracts a user's general preference from high-order feature interactions and sequential patterns of long history sequences; and (b) a sampling method to sample both the relatively long-term and short-term item histories. Experimental studies on two real-world datasets demonstrate considerable improvements for click-through rate (CTR) prediction accuracy relative to strong baseline models and also show the effectiveness of the student-model compression and acceleration for speed. [ABSTRACT FROM AUTHOR]

Aiming at the problem that traditional sentiment analysis methods cannot solve the problem of short video sentiment expression and the problem that existing multimodal sentiment analysis methods have low accuracy and poor interaction between different modal information, a multimodal sentiment analysis model with composite hierarchical fusion is established by studying multimodal sentiment analysis methods and combining temporal convolutional network (TCN) and Soft- attention mechanism. The model first dimensionally equalizes the text features, video facial features and audio features extracted from the video, and then fuses the obtained information features between different modalities in a composite manner (The model first fuses the unimodal information to obtain bimodal feature information, then fuses the obtained three bimodal information to obtain the final trimodal information. The final trimodal information and each unimodal information are fused to obtain the final multimodal sentiment feature information), and each fused feature information is then passed through the TCN network layer for sequence feature extraction, and the final multimodal feature information is used for sentiment classification after information filtering and feature dimensionality reduction through the screening attention mechanism to obtain the prediction results. Experiments on the datasets CMU-MOSI (CMU multimodal opinion level sentiment intensity) and CMU-MOSEI (CMU multimodal opinion sentiment and emotion intensity) show that the model can make full use of the interaction information between different modalities, so that the accuracy of multimodal sentiment analysis can be effectively improved. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Guangxi Normal University - Natural Science Edition is the property of Gai Kan Bian Wei Hui and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Click-through rate prediction is a critical task for computational advertising and recommendation systems, where the key challenge is to model feature interactions between different feature domains. At present, the main click-through rate prediction models model feature interactions in an implicit way, which leads to poor interpretation of the model, and the interaction between each pair of features may introduce noise into the model, thus limiting the predictive ability of the model. In response to the above problems, this paper proposes a click-through rate prediction model (GAIAN) based on the graph attention interactive aggregation network, which explicitly obtains cross features on the graph structure. Our specific method is to design a feature interactive selection mechanism to select cross features that are beneficial to model prediction, reducing model noise and reducing the risk of model overfitting. On this basis, the bilinear interaction function is integrated into the aggregation strategy of the graph neural network, and the fine-grained intersection features are extracted in a flexible and explicit way, which makes graph neural networks more suitable for modeling feature interactions and enhances the interpretability of the model. Compared with several other state-of-the-art models on the Criteo and Avazu datasets, the experimental results show the superiority of the model. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Computer Engineering & Applications is the property of Beijing Journal of Computer Engineering & Applications Journal Co Ltd. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)