Applied Intelligence最新文献

{"title":"DFF-HGNN: Dual-Feature Fusion Heterogeneous Graph Neural Network","authors":"Shengen Xue,&nbsp;Hua Duan,&nbsp;Yufei Zhao,&nbsp;Wei Fan","doi":"10.1007/s10489-025-06480-8","DOIUrl":"10.1007/s10489-025-06480-8","url":null,"abstract":"<div><p>Heterogeneous graph neural networks (HGNNs) have gained significant attention in deep learning due to their superior capability in processing heterogeneous graph data. However, existing HGNNs often fail to explicitly leverage relational information among nodes when utilizing the attribute information of nodes for graph representation learning, thus constraining their performance. To address this limitation, we introduce two approaches for utilizing relational information explicitly: a Relation-based Feature Enhancement Strategy (RFE-Strategy) for non-attributed heterogeneous graphs, and a Dual-Feature Fusion Heterogeneous Graph Neural Network (DFF-HGNN) for attributed heterogeneous graphs. The RFE-Strategy enhances HGNNs performance on non-attributed heterogeneous graphs through a three-step process: relational feature extraction, identity feature encoding, and feature enhancement. Meanwhile, DFF-HGNN integrates both attribute and relational features to effectively capture the heterogeneity and complexity of the graph, employing four components: separate pre-transformation, intra-type feature encoder, inter-type feature encoder, and embedding update encoder. Extensive experiments on multiple benchmark datasets demonstrate that the RFE-Strategy significantly improves the performance of HGNNs, while DFF-HGNN outperforms the state-of-the-art models.</p></div>","PeriodicalId":8041,"journal":{"name":"Applied Intelligence","volume":"55 7","pages":""},"PeriodicalIF":3.4,"publicationDate":"2025-04-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143835736","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A two-stage knowledge graph completion based on LLMs’ data augmentation and atrous spatial pyramid pooling","authors":"Na Zhou,&nbsp;Yuan Yuan,&nbsp;Lei Chen","doi":"10.1007/s10489-025-06556-5","DOIUrl":"10.1007/s10489-025-06556-5","url":null,"abstract":"<div><p>With the development of information technology, a large amount of unstructured and fragmented data is generated. Knowledge graphs can effectively integrate these fragmented data. Due to the difficulty of domain knowledge mining, knowledge graphs have problems of data sparseness and data missing. In addition, standard convolutional neural networks have limited capability in capturing feature interactions. To address data sparsity and the limitations of standard convolutional models, we propose DA-ARKGC, a two-stage knowledge graph completion model using wheat as a case study. In the first stage, to address the data sparsity problem, the rule mining data augmentation module (DA) based on large language models expands the wheat knowledge graph. In the second stage, the knowledge completion module (ARKGC) of the atrous spatial pyramid pooling with residual is introduced to achieve knowledge completion. The DA-ARKGC model was verified on the constructed wheat knowledge graph (Wheat_KG). Compared with ConvE, its MRR, Hits@1, Hits@3 and Hits@10 increased by 10% and 10.2%, 10.1% and 9.3%, respectively. In order to verify the effectiveness and generalization of the ARKGC module, experiments were conducted on the open-source datasets WN18 and FB15k. The results demonstrated that the model achieved optimal or sub-optimal performance compared with other baseline models.</p></div>","PeriodicalId":8041,"journal":{"name":"Applied Intelligence","volume":"55 7","pages":""},"PeriodicalIF":3.4,"publicationDate":"2025-04-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143835741","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Radial-based oversampling based on differential evolution for imbalanced data","authors":"Jun Chen,&nbsp;Meng Xia,&nbsp;Zhijie Wang","doi":"10.1007/s10489-025-06460-y","DOIUrl":"10.1007/s10489-025-06460-y","url":null,"abstract":"<div><p>Data imbalance remains a significant obstacle in many real-world applications. Although the Synthetic Minority Over-sampling Technique (SMOTE) and its variants are widely used to mitigate this issue, they often suffer from noise sensitivity, over-constraint, and over-generalization. In this paper, we introduce Radial-Based Oversampling based on Differential Evolution (DERBO), a novel algorithm that combines the global search strength of differential evolution (DE) with a radial basis function (RBF)-guided fitness strategy. By generating synthetic samples that are both diverse and closely aligned with the original minority distribution, DERBO effectively overcomes the limitations of existing methods. Extensive comparisons across 32 datasets against nine state-of-the-art imbalanced learning techniques demonstrate DERBO’s consistently superior performance, establishing it as a highly competitive and robust solution for addressing data imbalance.</p></div>","PeriodicalId":8041,"journal":{"name":"Applied Intelligence","volume":"55 7","pages":""},"PeriodicalIF":3.4,"publicationDate":"2025-04-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1007/s10489-025-06460-y.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143835742","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Correction to: Novel ensemble bagging-logistic regression algorithm for NoSQL database security","authors":"Anuradha Kanade,&nbsp;Amol D. Vibhute,&nbsp;Shantanu Kanade","doi":"10.1007/s10489-025-06456-8","DOIUrl":"10.1007/s10489-025-06456-8","url":null,"abstract":"","PeriodicalId":8041,"journal":{"name":"Applied Intelligence","volume":"55 7","pages":""},"PeriodicalIF":3.4,"publicationDate":"2025-04-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143835738","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Emp-EEK: generating empathetic responses via exemplars and external knowledge","authors":"Zikun Wang,&nbsp;Jing Li,&nbsp;Jinshui Lai,&nbsp;Donghong Han,&nbsp;Baiyou Qiao,&nbsp;Gang Wu","doi":"10.1007/s10489-025-06464-8","DOIUrl":"10.1007/s10489-025-06464-8","url":null,"abstract":"<div><p>Empathy plays a crucial role in human communication, and empathetic dialogue systems have garnered increasing research interest. However, accurately modeling and quantifying empathy remains challenging due to its inherently complex and multifaceted nature. Exemplar-based guidance has shown promise in enhancing empathetic response generation, yet existing approaches suffer from limitations such as noisy or irrelevant exemplars. To address these challenges, we propose <b>Emp-EEK</b>, an <b>Emp</b>athetic response generation model guided by <b>E</b>xemplars and <b>E</b>xternal <b>K</b>nowledge. Specifically, we employ a fine-tuned Dense Passage Retriever to jointly retrieve relevant exemplars based on both utterance-exemplar similarity and contextual proximity, ensuring more precise guidance for response generation. Furthermore, to enhance the system’s understanding of the speaker, we integrate external knowledge into the dialogue history, enriching contextual comprehension. To further elevate the level of empathy in responses, we introduce a multi-expert system that incorporates three independent decoders at the decoding stage. This design enables the model to effectively learn and capture the three key psychological mechanisms of empathetic communication: emotional reaction, interpretation, and exploration. Experimental results on the Empathetic-Dialogues dataset, evaluated through both automatic metrics and human judgments, demonstrate the effectiveness of our approach. Additionally, case studies analyzing the decoding process of different decoders highlight the strong interpretability of our model. Our code is publicly available at https://github.com/NEUWzk/Emp-EEK.</p></div>","PeriodicalId":8041,"journal":{"name":"Applied Intelligence","volume":"55 7","pages":""},"PeriodicalIF":3.4,"publicationDate":"2025-04-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143835737","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"An attributed network features learning method for over-indebtedness prediction","authors":"Fengzhang Chen,&nbsp;Zewei Long,&nbsp;Wei Wang,&nbsp;Kai Qi","doi":"10.1007/s10489-025-06538-7","DOIUrl":"10.1007/s10489-025-06538-7","url":null,"abstract":"<div><p>Over-indebtedness represents a financial anomaly and is widely regarded as an early indicator of financial distress. The recent advancements in machine learning techniques have enabled more accurate prediction of over-indebtedness. While existing forecasting models have contributed to mitigating the negative impacts of over-indebtedness, they typically fail to account for the influence of external factors on corporate debt decisions, which consequently limits their predictive accuracy. In response, this paper introduces a novel prediction model for over-indebtedness based on an attributed network feature learning approach for early warning. Building on previous research, the proposed model incorporates external information, such as interlocking directorate networks and product competition networks, as additional data sources for feature construction. By leveraging descriptive analytics and deep attributed network embedding methods, the model captures both individual and external features from social network data. To optimize the model’s performance, a generative classifier—specifically, the locally-weighted Expectation Maximization method for Naïve Bayes learning—is employed to handle the network-based features. The experimental results demonstrate that the proposed model performs effectively and offers valuable insights for integrating external information into financial prediction models.</p></div>","PeriodicalId":8041,"journal":{"name":"Applied Intelligence","volume":"55 7","pages":""},"PeriodicalIF":3.4,"publicationDate":"2025-04-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143835659","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Out-of-distribution detection using normalizing flows on the data manifold","authors":"Seyedeh Fatemeh Razavi,&nbsp;Mohammadmahdi Mehmanchi,&nbsp;Reshad Hosseini,&nbsp;Mostafa Tavassolipour","doi":"10.1007/s10489-025-06499-x","DOIUrl":"10.1007/s10489-025-06499-x","url":null,"abstract":"<div><p>Using the intuition that out-of-distribution data have lower likelihoods, a common approach for out-of-distribution detection involves estimating the underlying data distribution. Normalizing flows are likelihood-based generative models providing a tractable density estimation via dimension-preserving invertible transformations. Conventional normalizing flows are prone to fail in out-of-distribution detection, because of the well-known curse of dimensionality problem of the likelihood-based models. To solve the problem of likelihood-based models, some works try to modify likelihood for example by incorporating a data complexity measure. We observed that these modifications are still insufficient. According to the manifold hypothesis, real-world data often lie on a low-dimensional manifold. Therefore, we proceed by estimating the density on a low-dimensional manifold and calculating a distance from the manifold as a measure for out-of-distribution detection. We propose a powerful criterion that combines this measure with the modified likelihood measure based on data complexity. Extensive experimental results show that incorporating manifold learning while accounting for the estimation of data complexity improves the out-of-distribution detection ability of normalizing flows. This improvement is achieved without modifying the model structure or using auxiliary out-of-distribution data during training.</p></div>","PeriodicalId":8041,"journal":{"name":"Applied Intelligence","volume":"55 7","pages":""},"PeriodicalIF":3.4,"publicationDate":"2025-04-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143830760","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Dueling double deep Q-network-based stamping resources intelligent scheduling for automobile manufacturing in cloud manufacturing environment","authors":"Yanjuan Hu,&nbsp;Leiting Pan,&nbsp;Zhongxian Wen,&nbsp;You Zhou","doi":"10.1007/s10489-025-06524-z","DOIUrl":"10.1007/s10489-025-06524-z","url":null,"abstract":"<div><p>With the development of intelligent manufacturing, the automobile manufacturing industry has entered the\"AI +\"era, and the cloud manufacturing paradigm for the application of the automobile manufacturing industry is also in progress. As a key of automobile manufacturing, the stamping resources scheduling for automobile manufacturing (SRSAM) in the cloud manufacturing (CMfg) is characterized by unique domain-specific attributes concerning task architecture, the particularities of resource allocation, and the agility in transitioning between service types, which impedes the effective transference of classical manufacturing resource scheduling methodologies. Concurrently, the prevalent approaches to stamping scheduling concentrate predominantly on resources within the confines of stamping workshops and production lines, which are limited in scope. Such approaches are ill-suited for coping with the volatile and extensive resource landscape inherent to cloud manufacturing environments. To handle the above issues, this paper proposes to solve the SRSAM problem in CMfg with a novel scheduling model and intelligent scheduling method based on Dueling Double Deep Q-network (DDDQN). Firstly, we propose a stamping resource multi-objective scheduling model within the in-depth analysis of the SRSAM problem in CMfg and introduce a novel task structure to articulate the dependencies within the stamping tasks. Secondly, addressing the static and dynamic scheduling requirements, we construct a scheduling framework based on deep reinforcement learning, propose the strategy combination based on 5 resource selections and 12 task selections to generate <i>Agent</i>'s actions. Finally, integrating the proposed scheduling framework and model, the DDDQN algorithm is designed to solve the optimal scheduling scheme. Experimental results indicate that the proposed method consistently matches or exceeds other DRL algorithms, including proximal policy optimization (PPO), Q-learning, Deep Q-network (DQN), Double DQN (DDQN), and Dueling DQN in terms of scheduling performance and model training.</p></div>","PeriodicalId":8041,"journal":{"name":"Applied Intelligence","volume":"55 7","pages":""},"PeriodicalIF":3.4,"publicationDate":"2025-04-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143835660","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A multilevel attention network with sub-instructions for continuous vision-and-language navigation","authors":"Zongtao He,&nbsp;Liuyi Wang,&nbsp;Shu Li,&nbsp;Qingqing Yan,&nbsp;Chengju Liu,&nbsp;Qijun Chen","doi":"10.1007/s10489-025-06544-9","DOIUrl":"10.1007/s10489-025-06544-9","url":null,"abstract":"<div><p>The aim of vision-and-language navigation (VLN) is to develop agents that navigate mapless environments via linguistic and visual observations. Continuous VLN, which more accurately mirrors real-world conditions than its discrete counterpart does, faces unique challenges such as real-time execution, complex instruction understanding, and long sequence prediction. In this work, we introduce a multilevel instruction understanding mechanism and propose a multilevel attention network (MLANet) to address these challenges. Initially, we develop a nonlearning-based fast sub-instruction algorithm (FSA) to swiftly generate sub-instructions without the need for annotations, achieving a speed enhancement of 28 times over the previous methods. Subsequently, our multilevel attention (MLA) module dynamically integrates visual features with both high- and low-level linguistic semantics, forming multilevel global semantics to bolster the complex instruction understanding capabilities of the model. Finally, we introduce the peak attention loss (PAL), which enables the flexible and adaptive selection of the current sub-instruction, thereby improving accuracy and stability achieved for long trajectories by focusing on the relevant local semantics. Our experimental findings demonstrate that MLANet significantly outperforms the baselines and is applicable to real-world robots.</p></div>","PeriodicalId":8041,"journal":{"name":"Applied Intelligence","volume":"55 7","pages":""},"PeriodicalIF":3.4,"publicationDate":"2025-04-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143830702","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Measurement of adulteration in liquids by optical interferograms analysis and deep learning","authors":"P. Prieto-Cortés,&nbsp;E. López-Meléndez,&nbsp;R. I. Álvarez-Tamayo,&nbsp;A. Barcelata-Pinzón,&nbsp;L.D. Lara-Rodriguez","doi":"10.1007/s10489-025-06550-x","DOIUrl":"10.1007/s10489-025-06550-x","url":null,"abstract":"<div><p>We demonstrate the use of a proposed deep learning model to detect six different degrees of adulteration in alcoholic beverages by classifying interferograms captured through a dual aperture common-path interferometer (DACPI). The proposed two-arm convolutional neural network (TA-CNN) classifier is based on the extraction of linear and non-linear local features by principal components analysis (PCA) and t-distributed stochastic neighbor embedding (t-SNE), respectively. Then, the features of the reduced vectors are extracted individually with convolutional layers for the classification of three balanced sets of interferograms, with different initial calibration and external perturbation characteristics. In addition, an empirical study of the extracted vectors demonstrates the viability of our interferograms as candidates to be classified by the TA-CNN. The performance of the TA-CNN is compared with modern deep learning models adapted by transfer learning for this specific application. The results show a high average accuracy for all the deep models tested, both for separate and combined sets of 96% and 96.5%, respectively. The proposed TA-CNN is the best performance model, reaching an accuracy of 99.15% for the combined sets. Furthermore, an analysis based on the fast Fourier transform (FFT) corroborates the fact that the relevant information for the classification of interferograms lies in their phase. This approach represents a novel method in optical instrumentation without the use of traditional phase measurement interferometry, the need for highly optimized optical calibration, high-precision optical components, and the obtaining of interferograms datasets with the same DACPI setting up.</p></div>","PeriodicalId":8041,"journal":{"name":"Applied Intelligence","volume":"55 7","pages":""},"PeriodicalIF":3.4,"publicationDate":"2025-04-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143826603","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}