High-quality Representation Research Articles

Heterogeneous and clustering-enhanced personalized preference transfer for cross-domain recommendation

As a promising solution to alleviate the critical cold-start problem in recommendation systems, Cross-Domain Recommendation (CDR) aims to transfer users’ preferences from the source domain to the target domain. However, the negligence of semantic differences of heterogeneous relations when modeling user preferences and the non-consideration of a user’s common characteristics inferred from similar users when transferring the user’s preferences make recent CDR works cannot well handle the two core issues of CDR—‘what to transfer’ and ‘how to transfer’. To this end, we propose a novel heterogeneous and clustering-enhanced user preference transfer model for CDR (named HCCDR). To well address the first issue, the heterogeneous latent factor modeling component is firstly built to compute high-quality representations of users and items to be transferred based on the heterogeneous relations among users and items. Note that the heterogeneous relations with different semantics are processed with different models. Then, the clustering-enhanced preference transfer component well addresses the second issue by learning an effective personalized preference transfer function between two domains for a user, where the individual and common characteristics of the user are concurrently considered. Finally, the personalized recommendation component achieves the personalized recommendation in the target domain based on the transferred user embeddings calculated via the learned personalized preference transfer functions. Experimental results performed on large public datasets demonstrate that the proposed HCCDR markedly outperforms all baselines. In particular, HCCDR gains a maximum performance improvement of 12.69% (or 8.99%) for RMSE (or MAE) compared with the best baseline.

Meta-path infomax joint structure enhancement for multiplex network representation learning

Network representation learning has achieved significant success in homogeneous network data analysis in recent years. Nevertheless, they cannot be directly applied in multiplex networks. To overcome the characteristic of heterogeneity in multiplex networks, several emerging methods utilize the concept of meta-path to denote different types of relations and obtain the node representations for each type of meta-path individually. Despite the remarkable progress, there still exist two important issues in the previous approaches. First, the complementary information between different types of meta-paths that may make the representations more discriminative, is rarely investigated. Second, current studies generally learn multiplex node representations based on the original graph structure, while overlooking the latent relations between nodes. To address the aforementioned issues, in this paper, we propose a novel model with Meta-path Infomax joint Structure Enhancement (MISE) for multiplex network representations. Specifically, we first develop a meta-path infomax mechanism, which maximizes the mutual information between local and global meta-path representations, making the node representation contain more complementary information. Additionally, we propose a graph structure learning module that captures the implicit correlations between nodes to construct the latent graph structure. Such structure enhancement is a simple yet surprisingly effective technique to learn high-quality representations. We sufficiently evaluate the performance of our proposal on both supervised and unsupervised downstream tasks. Comprehensive experimental results show that our MISE achieves a promising boost in performance on a variety of real-world datasets for multiplex network representation learning.

Non-Contrastive Nearest Neighbor Identity-Guided Method for Unsupervised Object Re-Identification

Recently, self-paced contrastive learning has emerged as a promising method for unsupervised object re-identification. These methods generate pseudo labels, store centroid features in the memory bank, and periodically update them. However, affected by the performance of the clustering method, within each cluster exists inevitably noisy instances, and self-paced contrastive learning usually requires a large number of negative samples from various classes, where false-negative samples give rise to the class collision issue. These lead to performing incorrect model optimization. In this paper, we propose a non-contrastive nearest neighbor identity-guided (NNNI) method to overcome these challenges. The advantage of NNNI is to provide the model with a highly accurate prior. Specifically, this method relies on the random identity sampler commonly used in re-identification tasks to provide the network with a regression target of the nearest neighbors of the same identity within a mini-batch. It encodes more and more information in an iterative process through a Siamese network with an exponential moving average to train high-quality representations. NNNI alleviates the negative effects of noise instances and corrects class collision issues during training. Extensive experiments show that our method is effective on unsupervised object re-identification and achieves state-of-the-art performance on three large-scale person re-identification datasets and one large-scale vehicle re-identification dataset, which is competitive with even supervised methods.

Web Service Recommendation via Combining Bilinear Graph Representation and xDeepFM Quality Prediction

With the increasing number of Web services, how to provide developers with Web services that meet their Mashup requirements accurately and efficiently has become a challenging problem. Therefore, focusing on the problem of “recommending appropriate services to build high-quality Mashup applications”, this paper proposes a Web service recommendation method via combining bilinear graph attention representation and xDeepFM (eXtreme Deep Factorization Machine) quality prediction. This method is based on content and structure-oriented service func-tion classification and combines it with the service invocation prediction based on multi-dimensional quality attributes. Firstly, it uses the Word2Vec model to learn the latent semantic repre-sentations from service description documents. Then, it constructs the service relationship network according to tags and shared annotation relationships of Web services. Next, a bilinear aggre-gator is used to model the pairwise interactions between neighbor service nodes. Integrated with the traditional weighted sum ag-gregator, a bilinear graph neural network (BGNN) with stronger node representation ability is constructed. It exploits BGNN to calculate the representation of service nodes in the network and divides services into different functionality clusters. Finally, the high-quality representation results are combined with mul-ti-dimensional QoS attributes. Aiming at the Web services in the service cluster, it utilizes xDeepFM to model and mine the com-plex interactions between Web services’ features, and predict and rank the invocation scores of Web services. The experimental results on the real dataset of ProgrammableWeb show that com-pared with the other ten methods, the proposed approach has better performance in terms of Accuracy, Recall, F1, Logloss, and AUC, and has better performance in classification and recom-mendation.

A fairness-aware graph contrastive learning recommender framework for social tagging systems

Personalized recommendations for social tagging systems aim to deliver high-quality recommendations for users and annotate meaningful characteristics on items. However, compared to some state-of-the-art general recommender systems, current tag-aware recommender systems (TRS) have no advantage in either performance or efficiency. We discover that tagging information not only boost performance, but also promote fairness caused by biased data. In our work, we propose a fairness-aware graph contrastive learning framework named Tag-aware Graph Contrastive Learning (TAGCL). The design of its graph structure is composed of two bipartite graphs: user-tag graph and item-tag graph. Leveraging contrastive learning paradigm, our proposed TAGCL is able to encode stable and high-quality representations. In the fairness-aware learning, we jointly optimize the model through negative tag learning and TransT regularization. Extensive experiments on three public datasets and two sizable real-world datasets show that TAGCL outperforms some state-of-the-art general recommender methods and TRS in terms of accuracy of recommendations and dramatically reduces bias on some heavily skewed dataset. We also carry out comprehensive ablation studies to verify the validity of our proposed framework.

VIGA: A variational graph autoencoder model to infer user interest representations for recommendation

Learning representations of both user interests and item characteristics is essentially important for recommendation tasks. Although graph neural network-based methods have significant advantages, there are still two inherent limitations: (i) users or items are often modeled as single embedding points in the vector space, and (ii) these methods usually focus more on the graph structure of user–item interactions, rather than the attribute-related information and useful attribute interactions. These limitations make it difficult to model complex user interests and learn high-quality representations, thus fail to obtain high recommendation accuracy. To overcome these limitations, we propose a variational inference-based graph autoencoder (VIGA) model to explore a multivariate distribution over latent representations for recommendation. In brief, to identify both of the complex users' interests and their interactive behaviors, VIGA first encodes users and items into latent variables, and subjects them to a multivariate Gaussian distribution via a graph autoencoder framework based on Bayesian variational inference. Then a feature fusion layer is established for the exchange and integration of both the topological structure and nodes attributes in the user–item interaction graph. Experiments on three real-world datasets show that the proposed VIGA method significantly outperforms state-of-the-art recommendation methods on recommendation accuracy.

Enhancing Spam Message Classification and Detection Using Transformer-Based Embedding and Ensemble Learning.

Over the last decade, the Short Message Service (SMS) has become a primary communication channel. Nevertheless, its popularity has also given rise to the so-called SMS spam. These messages, i.e., spam, are annoying and potentially malicious by exposing SMS users to credential theft and data loss. To mitigate this persistent threat, we propose a new model for SMS spam detection based on pre-trained Transformers and Ensemble Learning. The proposed model uses a text embedding technique that builds on the recent advancements of the GPT-3 Transformer. This technique provides a high-quality representation that can improve detection results. In addition, we used an Ensemble Learning method where four machine learning models were grouped into one model that performed significantly better than its separate constituent parts. The experimental evaluation of the model was performed using the SMS Spam Collection Dataset. The obtained results showed a state-of-the-art performance that exceeded all previous works with an accuracy that reached 99.91%.

Heterogeneous Graph Purification Network: Purifying Noisy Heterogeneity without Metapaths

Heterogeneous graph neural networks (HGNNs) deliver the powerful capability to model many complex systems in real-world scenarios by embedding rich structural and semantic information of a heterogeneous graph into low-dimensional representations. However, existing HGNNs encounter great difficulty in balancing the ability to avoid artificial metapaths with resisting structural and informational noise in a heterogeneous graph. In this paper, we propose a novel framework called Heterogeneous Graph Purification Network (HGPN) which aims to solve such dilemma by adaptively purifying the noisy heterogeneity. Specifically, instead of relying on artificial metapaths, HGPN models heterogeneity by subgraph decomposition and adopts inter-subgraph and intra-subgraph aggregation methods. HGPN can learn to purify noisy edges based on semantic information with a parallel heterogeneous structure purification mechanism. Besides, we design a neighborhood-related dynamic residual update method, a type-specific normalization module and cluster-aware loss to help all types of node achieve high-quality representations and maintain feature distribution while preventing feature over-mixing problems. Extensive experiments are conducted on four common heterogeneous graph datasets, and results show that our approach outperforms all existing methods and achieves state-of-the-art performances consistently among all the datasets.

Does Spanish knowledge contribute to accurate English word spelling in adult bilinguals?

AbstractCorrectly spelling an English word requires a high-quality orthographic representation. When faced with spelling a complex word without a high-quality representation, spellers often rely on other knowledge sources (e.g., incomplete stored orthographic forms, phonological to orthographic relationships) to spell it. For bilinguals, another potentially facilitative source is knowledge of a word's lexical and sublexical representations in another language. In the current study we considered simultaneous effects of word-level (e.g., frequency, cognate status) and person-level (e.g., English spelling skill, prompting, bilingual status) predictors on college students’ complex English word spelling. Monolinguals (English; n = 42) significantly outperformed bilinguals (Spanish and English; n = 76) on non-cognate spelling; no group differences emerged for cognate spelling accuracy. Within bilinguals, significantly higher spelling performance on cognates compared to non-cognates suggests cognate facilitation, with no prompting effects. Findings expand an interdisciplinary framework of understanding bilinguals’ activation and use of cross-linguistic representations in spelling.

A novel fault detection method for rotating machinery based on self-supervised contrastive representations

In many industrial applications, only healthy condition monitoring data is available for rotating machines that are newly put into operation or have not met the onset of degradation, i.e. the occurrence of incipient failures, after a long-time run. Fault detection based on only the healthy condition data can be a challenging one-class classification problem, whose key is to extract high-quality representations of healthy condition which can be distinguished from those of faulty conditions to the maximum extent. However, the widely adopted hand-crafted representations composed of statistical metric values and representations extracted by data reconstruction methods cannot guarantee to be faithful to the fault detection task, because their targets are indirect regarding the extraction of discriminative representations. In this work, we propose a novel fault detection framework based on i) the self-supervised representations extracted from the vibration signals, the most widely used for rotating machine monitoring, of healthy condition using contrastive learning and ii) one-class classifiers. A negative sequence construction strategy is developed, which allows the extraction of intrinsic representations of healthy condition data using contrastive learning by distinguishing a data sample from its augmentations. The extracted representation is sensitive to the modifications of healthy data and, therefore, is more faithful to the one-class classification task. The performance of the proposed framework is verified using experimental data of bearings and gears, which are key vulnerable components of rotating machines. Its advantage is shown by comparing it with state-of-the-art methods based on other unsupervised representations and one-class classifiers.

L2MM: Learning to Map Matching with Deep Models for Low-Quality GPS Trajectory Data

Map matching is a fundamental research topic with the objective of aligning GPS trajectories to paths on the road network. However, existing models fail to achieve satisfactory performance for low-quality (i.e., noisy, low-frequency, and non-uniform) trajectory data. To this end, we propose a general and robust deep learning-based model, L2MM , to tackle these issues at all. First, high-quality representations of low-quality trajectories are learned by two representation enhancement methods, i.e., enhancement with high-frequency trajectories and enhancement with the data distribution . The former employs high-frequency trajectories to enhance the expressive capability of representations, while the latter regularizes the representation distribution over the latent space to improve the generalization ability of representations. Secondly, to embrace more heuristic clues, typical mobility patterns are recognized in the latent space and further incorporated into the map matching task. Finally, based on the available representations and patterns, a mapping from trajectories to corresponding paths is constructed through a joint optimization method. Extensive experiments are conducted based on a range of datasets, which demonstrate the superiority of L2MM and validate the significance of high-quality representations as well as mobility patterns.

User Cold-Start Recommendation via Inductive Heterogeneous Graph Neural Network

Recently, user cold-start recommendations have attracted a lot of attention from industry and academia. In user cold-start recommendation systems, the user attribute information is often used by existing approaches to learn user preferences due to the unavailability of user action data. However, most existing recommendation methods often ignore the sparsity of user attributes in cold-start recommendation systems. To tackle this limitation, this article proposes a novel Inductive Heterogeneous Graph Neural Network (IHGNN) model, which utilizes the relational information in user cold-start recommendation systems to alleviate the sparsity of user attributes. Our model converts new users, items, and associated multimodal information into a Modality-aware Heterogeneous Graph (M-HG) that preserves the rich and heterogeneous relationship information among them. Specifically, to utilize rich and heterogeneous relational information in an M-HG for enriching the sparse attribute information of new users, we design a strategy based on random walk operations to collect associated neighbors of new users by multiple times sampling operation. Then, a well-designed multiple hierarchical attention aggregation model consisting of the intra- and inter-type attention aggregating module is proposed, focusing on useful connected neighbors and neglecting meaningless and noisy connected neighbors to generate high-quality representations for user cold-start recommendations. Experimental results on three real datasets demonstrate that the IHGNN outperforms the state-of-the-art baselines.

Improving drug-drug interactions prediction with interpretability via meta-path-based information fusion.

Drug-drug interactions (DDIs) are compound effects when patients take two or more drugs at the same time, which may weaken the efficacy of drugs or cause unexpected side effects. Thus, accurately predicting DDIs is of great significance for the drug development and the drug safety surveillance. Although many methods have been proposed for the task, the biological knowledge related to DDIs is not fully utilized and the complex semantics among drug-related biological entities are not effectively captured in existing methods, leading to suboptimal performance. Moreover, the lack of interpretability for the predicted results also limits the wide application of existing methods for DDIs prediction. In this study, we propose a novel framework for predicting DDIs with interpretability. Specifically, we construct a heterogeneous information network (HIN) by explicitly utilizing the biological knowledge related to the procedure of inducing DDIs. To capture the complex semantics in HIN, a meta-path-based information fusion mechanism is proposed to learn high-quality representations of drugs. In addition, an attention mechanism is designed to combine semantic information obtained from meta-paths with different lengths to obtain final representations of drugs for DDIs prediction. Comprehensive experiments are conducted on 2410 approved drugs, and the results of predictive performance comparison show that our proposed framework outperforms selected representative baselines on the task of DDIs prediction. The results of ablation study and cold-start scenario indicate that the meta-path-based information fusion mechanism red is beneficial for capturing the complex semantics among drug-related biological entities. Moreover, the results of case study demonstrate that the designed attention mechanism is able to provide partial interpretability for the predicted DDIs. Therefore, the proposed method will be a feasible solution to the task of predicting DDIs.

A bidirectional trajectory contrastive learning model for driving intention prediction

AbstractDriving intention prediction with trajectory data of surrounding vehicles is critical to advanced driver assistance system for improving the accuracy of decision-making. Previous works mostly focused on trajectory representation based on supervised manners. However, learning generalized and high-quality representations from unlabeled data remains a very challenging task. In this paper, we propose a self-supervised bidirectional trajectory contrastive learning (BTCL) model that learns generalized trajectory representation to improve the performance of the driving intention prediction task. Different trajectory data augmentation strategies and a cross-view trajectory prediction task are constructed jointly as pretext task of contrastive learning. The pretext task can maximize the similarity among different augmentations of the same sample while minimizing similarity among augmentations of different samples. It can not only learn the high-quality representation of trajectory without labeled information but also improve the adversarial attacks on BTCL. Moreover, considering the vehicle trajectory forward and backward follows the same social norms and driving behavior constraints. A bidirectional trajectory contrastive learning module is built to gain more positive samples that further increasing the prediction accuracy in downstream tasks and transfer ability of the model. Experimental results demonstrate that BTCL is competitive with the state-of-the-art, especially for adversarial attack and transfer learning tasks, on real-world HighD and NGSIM datasets.

Quantitative Evaluation of Bathymetric LiDAR Sensors and Acquisition Approaches in Lærdal River in Norway

The development of bathymetric LiDAR technology has contributed significantly to both the quality and quantity of river bathymetry data. Although several bathymetric LiDAR sensors are available today, studies that evaluate the performance of the different bathymetric LiDAR sensors comparatively are still lacking. This study evaluates the performance of three bathymetric LiDAR sensors, CZMIL Supernova, Riegl VQ880-G, and Riegl VQ840-G, used with different acquisition approaches, in mapping Lærdal River bathymetry in Norway. The performance was evaluated based on comparing the sensors against a multibeam echosounder (MBES), a terrestrial laser scanner (TLS), and by an intercomparison between the individual sensors. The comparison was completed by comparing point clouds from the instruments and through the comparison of DEMs created from the point clouds. For the comparison against the MBES, the results show that the median residuals range between 3 to 13 cm, while against the TLS the median residuals range between 0 to 5 cm. The comparison of the CZMIL sensor against the two Riegl sensors shows median residuals of around 12 cm where the CZMIL map is shallower against the VQ880-G and deeper against the VQ840-G sensor. For the two Riegl sensors, the results show a median difference of 2 cm with the VQ880-G map deeper. We do observe that areas with high residuals are linked to river features such as large substrate variability, steep banks, and whitewater/turbulent flow. The study shows that all the LiDAR instruments provide high-quality representations of the river geometry and create a solid foundation for planning, modelling, or other work in rivers where detailed bathymetry is needed.

Heterogeneous Graph Contrastive Learning Network for Personalized Micro-Video Recommendation

Personalized micro-video recommendation has attracted a lot of research attention with the growing popularity of micro-video sharing platforms. Many efforts have been made to consider micro-video recommendation as a matching task and shown promising performance, while they only focus on simple features or multi-modal attribute information. Recently, Graph Neural Networks (GNNs) have been employed in many recommendation tasks and achieved impressive success. However, these GNN-based methods may suffer from the following limitations: (1) fail to capture the heterogeneity of nodes in uservideo bipartite graphs; (2) ignore the non-local (global) semantic correlation information remained in heterogeneous graphs. In this paper, we present a novel approach, Heterogeneous Graph Contrastive Learning Network (HGCL), for personalized microvideo recommendation. To consider heterogeneity in user-video bipartite graphs, we first introduce a heterogeneous graph encoder network for a high-quality representation learning of users and micro-videos. Specifically, we design a random surfing model to generate node-type specific homogeneous graphs to preserve the heterogeneity. Then we propose a graph contrastive learning framework to achieve representation learning on each node-type specific homogeneous graph by maximizing the mutual information between local patches of a graph and the global representation of the entire graph. Finally, a type-crossing objective function is proposed to jointly integrate the node embeddings from different node types to facilitate high-quality representation learning. Experimental results on real-world datasets in the micro-video recommendation task validate the performance of our method, compared with state-of-the-art baseline algorithms.

Critical Relation Path Aggregation-Based Industrial Control Component Exploitable Vulnerability Reasoning

With the growing discovery of exposed vulnerabilities in the Industrial Control Components (ICCs), identification of the exploitable ones is urgent for Industrial Control System (ICS) administrators to proactively forecast potential threats. However, it is not a trivial task due to the complexity of the multi-source heterogeneous data and the lack of automatic analysis methods. To address these challenges, we propose an exploitability reasoning method based on the ICC-Vulnerability Knowledge Graph (KG) in which relation paths contain abundant potential evidence to support the reasoning. The reasoning task in this work refers to determining whether a specific relation is valid between an attacker entity and a possible exploitable vulnerability entity with the help of a collective of the critical paths. The proposed method consists of three primary building blocks: KG construction, relation path representation, and query relation reasoning. A security-oriented ontology combines exploit modeling, which provides a guideline for the integration of the scattered knowledge while constructing the KG. We emphasize the role of the aggregation of the attention mechanism in representation learning and ultimate reasoning. In order to acquire a high-quality representation, the entity and relation embeddings take advantage of their local structure and related semantics. Some critical paths are assigned corresponding attentive weights and then they are aggregated for the determination of the query relation validity. In particular, similarity calculation is introduced into a critical path selection algorithm, which improves search and reasoning performance. Meanwhile, the proposed algorithm avoids redundant paths between the given pairs of entities. Experimental results show that the proposed method outperforms the state-of-the-art ones in the aspects of embedding quality and query relation reasoning accuracy.

Parameter uncertainty modeling for multiobjective robust control design. Application to a temperature control system in a proton exchange membrane fuel cell

Advanced control systems are tuned using dynamic models and optimization techniques. This approach frequently involves satisfying multiple conflicting objectives. Tuning robust controllers requires considering a framework that represents the system uncertainties, and its definition is not a trivial task. When dealing with a nonlinear model with many parameters, a high-quality representation requires a massive sampling of variations. In many cases, this represents an inaccessible computational cost for the optimization process.This work presents a new methodology for parameter uncertainty modeling that is oriented to tuning robust controllers based on multiobjective optimization techniques. The uncertainty modeling formulated represents a feasible computational cost and leads to robust solutions without attributing excessive conservatism. The novelty of this process consists in using the multiobjective space to define a set of scenarios with highly representative properties of the global uncertainty framework that formulate the control problem under a predefined minimization strategy.To demonstrate the effectiveness of this methodology, we present a temperature control design in a micro-CHP system under worst-case minimization. Based on the results, particular interest is given to verifying the appropriate formulation of the uncertainty modeling, which represents a 92.8% reduction of the computational cost involved in solving the robust optimization problem under a global uncertainty framework.

Consensus and complementary regularized non-negative matrix factorization for multi-view image clustering

Multi-view clustering (MVC) aims to classify objects using multi-view data. Many existing MVC methods explore the consensus agreement among views or enhance the complementary features from multiple views in order to improve the clustering quality. However, multi-view data, especially multi-view image data, contains much more complex and meaningful information, which not only describes intrinsic common properties of one object, but also implies rich unique features from different views. Both of them are important and useful for MVC. Though the widely-used non-negative matrix factorization (NMF) based MVC methods can address these two kinds of information, the qualities of learned representations are not well, which limits the clustering performance. To solve this problem, motivated by the ideas of feature selection and graph regularization, a novel NMF based unified representation learning framework is presented with the integration of two specifically designed graph regularization terms in order to obtain a high-quality representation for multi-view data: An unified NMF based optimization problem is formulated to learn the consensus and complementary representations simultaneously and an alternating optimization algorithm is designed to solve this non-convex optimization problem. Then, these learned representations are efficiently fused to create an integrated graph representation for the final clustering. Extensive experiments on six real image datasets demonstrate the superior performance of the proposed method compared with other eleven competitive MVC methods.

Multi-Level Knowledge-Aware Contrastive Learning Network for Personalized Recipe Recommendation

Personalized recipe recommendation is attracting more and more attention, which can help people make choices from the exploding growth of online food information. Unlike other recommendation tasks, the target of recipe recommendation is a non-atomic item, so attribute information is especially important for the representation of recipes. However, traditional collaborative filtering or content-based recipe recommendation methods tend to focus more on user–recipe interaction information and ignore higher-order semantic and structural information. Recently, graph neural networks (GNNs)-based recommendation methods provided new ideas for recipe recommendation, but there was a problem of sparsity of supervised signals caused by the long-tailed distribution of heterogeneous graph entities. How to construct high-quality representations of users and recipes becomes a new challenge for personalized recipe recommendation. In this paper, we propose a new method, a multi-level knowledge-aware contrastive learning network (MKCLN) for personalized recipe recommendation. Compared with traditional comparative learning, we design a multi-level view to satisfy the requirement of fine-grained representation of users and recipes, and use multiple knowledge-aware aggregation methods for node fusion to finally make recommendations. Specifically, the local-level includes two views, interaction view and semantic view, which mine collaborative information and semantic information for high-quality representation of nodes. The global-level learns node embedding by capturing higher-order structural information and semantic information through a network structure view. Then, a kind of self-supervised cross-view contrastive learning is invoked to make the information of multiple views collaboratively supervise each other to learn fine-grained node embeddings. Finally, the recipes that satisfy personalized preferences are recommended to users by joint training and model prediction functions. In this study, we conduct experiments on two real recipe datasets, and the experimental results demonstrate the effectiveness and advancement of MKCLN.