Engineering Applications of Artificial Intelligence最新文献_第2页

Automated acquisition and analysis of illegal fund accounts in gambling websites

IF 7.5 2区计算机科学 Q1 AUTOMATION & CONTROL SYSTEMS

Engineering Applications of Artificial Intelligence

Pub Date : 2025-02-25 DOI: 10.1016/j.engappai.2025.110326

Shenao Zheng , Yanan Cheng, Guoying Sun , Zhaoxin Zhang

While the rapid growth of the internet has made life more convenient, it has also made it easier for people to access online gambling sites. This has led to an increase in fraud, resulting in significant financial losses and serious societal issues. However, the abundance of gambling sites, diverse development frameworks, and complex identity verification methods pose significant challenges to gaining comprehensive access to their illegal funds. To address online gambling fraud and assist law enforcement in cutting off their funding chains, we investigate gambling websites to uncover information about the physical funding accounts used by criminal groups. Given the extensive scale of these websites, we propose an auto-registration framework based on You Only Look Once version 4 (YOLOv4) to automate account registration and retrieve illegal fund account details. Additionally, we construct a dataset of user interface (UI) elements from 17 types of gambling websites for model training. The YOLOv4 model achieves an F1-score of 0.84 and a mean Average Precision (mAP) of 83.96%. The proposed framework achieves a registration success rate of 60.58%. After extracting data from numerous gambling websites in seven batches, we identify 7496 entity account details and 47 payment methods, analyze the wealth of entity information, and highlight six new payment methods. Finally, by identifying recurring illegal fund accounts across multiple domains, we confirm 23 criminal gangs, providing substantial support to law enforcement agencies in combating online gambling-related crimes.

互联网的迅猛发展在为人们的生活带来便利的同时，也使人们更容易访问在线赌博网站。这导致了欺诈行为的增加，造成了巨大的经济损失和严重的社会问题。然而，大量的赌博网站、多样化的开发框架和复杂的身份验证方法给全面获取非法资金带来了巨大挑战。为了应对网络赌博欺诈，协助执法部门切断其资金链，我们对赌博网站进行调查，以发现犯罪集团使用的实际资金账户信息。鉴于这些网站规模庞大，我们提出了一个基于 You Only Look Once version 4（YOLOv4）的自动注册框架，以自动注册账户并获取非法资金账户的详细信息。此外，我们还构建了一个包含 17 种赌博网站用户界面（UI）元素的数据集，用于模型训练。YOLOv4 模型的 F1 分数为 0.84，平均精度 (mAP) 为 83.96%。拟议框架的注册成功率为 60.58%。在分七批次从众多赌博网站提取数据后，我们识别出 7496 个实体账户详情和 47 种支付方式，分析了大量实体信息，并突出显示了六种新的支付方式。最后，通过识别多个域名中重复出现的非法资金账户，我们确认了 23 个犯罪团伙，为执法部门打击网络赌博相关犯罪提供了有力支持。

{"title":"Automated acquisition and analysis of illegal fund accounts in gambling websites","authors":"Shenao Zheng , Yanan Cheng, Guoying Sun , Zhaoxin Zhang","doi":"10.1016/j.engappai.2025.110326","DOIUrl":"10.1016/j.engappai.2025.110326","url":null,"abstract":"<div><div>While the rapid growth of the internet has made life more convenient, it has also made it easier for people to access online gambling sites. This has led to an increase in fraud, resulting in significant financial losses and serious societal issues. However, the abundance of gambling sites, diverse development frameworks, and complex identity verification methods pose significant challenges to gaining comprehensive access to their illegal funds. To address online gambling fraud and assist law enforcement in cutting off their funding chains, we investigate gambling websites to uncover information about the physical funding accounts used by criminal groups. Given the extensive scale of these websites, we propose an auto-registration framework based on You Only Look Once version 4 (YOLOv4) to automate account registration and retrieve illegal fund account details. Additionally, we construct a dataset of user interface (UI) elements from 17 types of gambling websites for model training. The YOLOv4 model achieves an F1-score of 0.84 and a mean Average Precision (mAP) of 83.96%. The proposed framework achieves a registration success rate of 60.58%. After extracting data from numerous gambling websites in seven batches, we identify 7496 entity account details and 47 payment methods, analyze the wealth of entity information, and highlight six new payment methods. Finally, by identifying recurring illegal fund accounts across multiple domains, we confirm 23 criminal gangs, providing substantial support to law enforcement agencies in combating online gambling-related crimes.</div></div>","PeriodicalId":50523,"journal":{"name":"Engineering Applications of Artificial Intelligence","volume":"147 ","pages":"Article 110326"},"PeriodicalIF":7.5,"publicationDate":"2025-02-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143479003","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}

引用次数: 0

Intelligent application of interactive scale transformer for fine grained feature extraction in sheep

IF 7.5 2区计算机科学 Q1 AUTOMATION & CONTROL SYSTEMS

Engineering Applications of Artificial Intelligence

Pub Date : 2025-02-25 DOI: 10.1016/j.engappai.2025.110300

Xingran Guo , Haizheng Yu , Hong Bian , Wenrong Li , Xueying Liao , Yongqi Zhu

In modern animal husbandry, artificial intelligence helps accurately manage individual sheep. However, it is difficult to recognize the sheep’s facial features and capture the nuances. It is not easy to extract the fine-grained features of a sheep’s face because the traditional vision transformer cannot realize the effective embedding of the interaction scale. To address this problem, we propose a novel sheep Transformer tool called SheepFormer. This model comprises components such as the Interactive Scale Embedded Images Block (ISEI), Patch Short Long Distance Attention Module (PSLDA), Dynamic Relative Position Offset (DRPO), and Transformer Neck and Head (TNH). These components are designed to embed features at multiple scales, fuse long and short-distance self-attention, adaptively handle relative position offsets for various group sizes, and introduce a prediction head to detect fine-grained facial targets in sheep at different scales. SheepFormer integrates Residual Attention to seek dense facial features in sheep and utilizes a Transformer Head to replace the traditional head, exploring the predictive potential of self-attention mechanisms in sheep faces. Experimental results demonstrate a 7.1% improvement in average precision (AP) for sheep face detection compared to Collaborative DEtection TRansformer (CO-DETR) and an 11.12% enhancement in accurate classification of sheep identity document (ID) compared to Shifted Windows Transformer (Swin Transformer). This study demonstrates that SheepFormer can extract fine-grained facial features of sheep, which promotes the advancement of high-precision sheep individual recognition and provides a guide for recognizing kinship.

{"title":"Intelligent application of interactive scale transformer for fine grained feature extraction in sheep","authors":"Xingran Guo , Haizheng Yu , Hong Bian , Wenrong Li , Xueying Liao , Yongqi Zhu","doi":"10.1016/j.engappai.2025.110300","DOIUrl":"10.1016/j.engappai.2025.110300","url":null,"abstract":"<div><div>In modern animal husbandry, artificial intelligence helps accurately manage individual sheep. However, it is difficult to recognize the sheep’s facial features and capture the nuances. It is not easy to extract the fine-grained features of a sheep’s face because the traditional vision transformer cannot realize the effective embedding of the interaction scale. To address this problem, we propose a novel sheep Transformer tool called <strong>SheepFormer</strong>. This model comprises components such as the Interactive Scale Embedded Images Block (<strong>ISEI</strong>), Patch Short Long Distance Attention Module (<strong>PSLDA</strong>), Dynamic Relative Position Offset (<strong>DRPO</strong>), and Transformer Neck and Head (<strong>TNH</strong>). These components are designed to embed features at multiple scales, fuse long and short-distance self-attention, adaptively handle relative position offsets for various group sizes, and introduce a prediction head to detect fine-grained facial targets in sheep at different scales. SheepFormer integrates Residual Attention to seek dense facial features in sheep and utilizes a Transformer Head to replace the traditional head, exploring the predictive potential of self-attention mechanisms in sheep faces. Experimental results demonstrate a 7.1% improvement in average precision (AP) for sheep face detection compared to Collaborative DEtection TRansformer (CO-DETR) and an 11.12% enhancement in accurate classification of sheep identity document (ID) compared to Shifted Windows Transformer (Swin Transformer). This study demonstrates that SheepFormer can extract fine-grained facial features of sheep, which promotes the advancement of high-precision sheep individual recognition and provides a guide for recognizing kinship.</div></div>","PeriodicalId":50523,"journal":{"name":"Engineering Applications of Artificial Intelligence","volume":"147 ","pages":"Article 110300"},"PeriodicalIF":7.5,"publicationDate":"2025-02-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143478995","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}

引用次数: 0

Regression loss-assisted conditional style generative adversarial network for virtual sample generation with small data in soft sensing

IF 7.5 2区计算机科学 Q1 AUTOMATION & CONTROL SYSTEMS

Engineering Applications of Artificial Intelligence

Pub Date : 2025-02-25 DOI: 10.1016/j.engappai.2025.110306

Xue-Yu Zhang , Qun-Xiong Zhu , Wei Ke , Yan-Lin He , Ming-Qing Zhang , Yuan Xu

Existing methods that extend virtual sample pools to address small sample problem caused by sample atypicality and uneven distribution often overlook data sparsity and inverse sample generation challenges, which limits the accuracy of subsequent modeling. To address above problem, we propose a novel regression-assisted conditional style generative adversarial network (RAC-StyleGAN). The proposed method leverages the strengths of StyleGAN in latent space mapping to enhance data diversity and granularity, while incorporating regression-assisted conditions to improve modeling performance. Specifically, RAC-StyleGAN utilizes kernel density estimation and radial basis function interpolation to ensure that the generated output variables are uniformly distributed. Based on the principle of inverse transformation, the interpolated output variables are then used as conditional inputs for the StyleGAN model, generating virtual input variables that faithfully reflect the marginal distribution of the original data. Furthermore, to preserve the complex nonlinear relationships between input and output variables, RAC-StyleGAN integrates a regression loss strategy based on empirical risk minimization into the StyleGAN framework. By fine-tuning the generation process, the soft-sensing model effectively captures the nonlinear mapping between inputs and outputs. Experimental validations on synthetic nonlinear functions, University of California Irvine machine learning (UCI) datasets, and a real-world purified terephthalic acid (PTA) solvent system demonstrate that RAC-StyleGAN effectively generates high-quality virtual samples, significantly enhancing the modeling performance.

{"title":"Regression loss-assisted conditional style generative adversarial network for virtual sample generation with small data in soft sensing","authors":"Xue-Yu Zhang , Qun-Xiong Zhu , Wei Ke , Yan-Lin He , Ming-Qing Zhang , Yuan Xu","doi":"10.1016/j.engappai.2025.110306","DOIUrl":"10.1016/j.engappai.2025.110306","url":null,"abstract":"<div><div>Existing methods that extend virtual sample pools to address small sample problem caused by sample atypicality and uneven distribution often overlook data sparsity and inverse sample generation challenges, which limits the accuracy of subsequent modeling. To address above problem, we propose a novel regression-assisted conditional style generative adversarial network (RAC-StyleGAN). The proposed method leverages the strengths of StyleGAN in latent space mapping to enhance data diversity and granularity, while incorporating regression-assisted conditions to improve modeling performance. Specifically, RAC-StyleGAN utilizes kernel density estimation and radial basis function interpolation to ensure that the generated output variables are uniformly distributed. Based on the principle of inverse transformation, the interpolated output variables are then used as conditional inputs for the StyleGAN model, generating virtual input variables that faithfully reflect the marginal distribution of the original data. Furthermore, to preserve the complex nonlinear relationships between input and output variables, RAC-StyleGAN integrates a regression loss strategy based on empirical risk minimization into the StyleGAN framework. By fine-tuning the generation process, the soft-sensing model effectively captures the nonlinear mapping between inputs and outputs. Experimental validations on synthetic nonlinear functions, University of California Irvine machine learning (UCI) datasets, and a real-world purified terephthalic acid (PTA) solvent system demonstrate that RAC-StyleGAN effectively generates high-quality virtual samples, significantly enhancing the modeling performance.</div></div>","PeriodicalId":50523,"journal":{"name":"Engineering Applications of Artificial Intelligence","volume":"147 ","pages":"Article 110306"},"PeriodicalIF":7.5,"publicationDate":"2025-02-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143478996","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}

引用次数: 0

A three layer stacked multimodel transfer learning approach for deep feature extraction from Chest Radiographic images for the classification of COVID-19

IF 7.5 2区计算机科学 Q1 AUTOMATION & CONTROL SYSTEMS

Engineering Applications of Artificial Intelligence

Pub Date : 2025-02-25 DOI: 10.1016/j.engappai.2025.110241

Baijnath Kaushik, Akshma Chadha, Abhigya Mahajan, Malvika Ashok

COVID-19 has had a profound impact on global health, targeting the human respiratory system and causing significant disruptions to human life. To aid in effective diagnosis and classification, numerous machine learning and deep learning models have been developed to analyse limited Chest Radiographic and Computed Tomography images. In this study, we propose a three-layer stacked multimodal approach for deep feature extraction from a high volume of COVID-19 Chest Radiographic images. The proposed model utilizes eight transfer learning models pre-trained on the ImageNet dataset, evaluated based on key performance metrics such as accuracy, precision, and recall. The unique stacking model integrates the outputs of these transfer learning models, extracting deep features through a three-layer architecture. These features are flattened, concatenated, and passed through seven deep dense layers with varying kernel and bias dimensions to achieve optimal classification performance. The approach was applied to a large COVID-19 Chest Radiographic images dataset, achieving the highest accuracy (95.79%), precision (95.44%), and recall (96.65%) when compared to state-of-the-art models. This study demonstrates the effectiveness of leveraging a stacked transfer learning multimodal framework for COVID-19 diagnosis. The proposed method not only ensures high accuracy but also provides a computationally efficient solution for analysing large-scale radiographic datasets, positioning it as a robust tool for aiding in the early detection and classification of COVID-19.

{"title":"A three layer stacked multimodel transfer learning approach for deep feature extraction from Chest Radiographic images for the classification of COVID-19","authors":"Baijnath Kaushik, Akshma Chadha, Abhigya Mahajan, Malvika Ashok","doi":"10.1016/j.engappai.2025.110241","DOIUrl":"10.1016/j.engappai.2025.110241","url":null,"abstract":"<div><div>COVID-19 has had a profound impact on global health, targeting the human respiratory system and causing significant disruptions to human life. To aid in effective diagnosis and classification, numerous machine learning and deep learning models have been developed to analyse limited Chest Radiographic and Computed Tomography images. In this study, we propose a three-layer stacked multimodal approach for deep feature extraction from a high volume of COVID-19 Chest Radiographic images. The proposed model utilizes eight transfer learning models pre-trained on the ImageNet dataset, evaluated based on key performance metrics such as accuracy, precision, and recall. The unique stacking model integrates the outputs of these transfer learning models, extracting deep features through a three-layer architecture. These features are flattened, concatenated, and passed through seven deep dense layers with varying kernel and bias dimensions to achieve optimal classification performance. The approach was applied to a large COVID-19 Chest Radiographic images dataset, achieving the highest accuracy (95.79%), precision (95.44%), and recall (96.65%) when compared to state-of-the-art models. This study demonstrates the effectiveness of leveraging a stacked transfer learning multimodal framework for COVID-19 diagnosis. The proposed method not only ensures high accuracy but also provides a computationally efficient solution for analysing large-scale radiographic datasets, positioning it as a robust tool for aiding in the early detection and classification of COVID-19.</div></div>","PeriodicalId":50523,"journal":{"name":"Engineering Applications of Artificial Intelligence","volume":"147 ","pages":"Article 110241"},"PeriodicalIF":7.5,"publicationDate":"2025-02-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143489018","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}

引用次数: 0

Rule training by score-based supervised contrastive learning for sketch explanation

IF 7.5 2区计算机科学 Q1 AUTOMATION & CONTROL SYSTEMS

Engineering Applications of Artificial Intelligence

Pub Date : 2025-02-25 DOI: 10.1016/j.engappai.2025.110310

Tae-Gyun Lee, Jang-Hee Yoo

This paper presents a novel approach to explain scoring results of infant visual-motor integration sketches utilized in developmental tests by training predefined rules for each test item. To address the performance issues caused by limited data, we employ a pre-trained model that uses supervised contrastive learning based on item scores. To ensure effective training, a memory bank structure is proposed to accumulate diverse embeddings over multiple iterations and prevent the encoder that processes item information from being trained to prevent collapsing in the Siamese network. Experiments demonstrate that the proposed method improves performance in both score and rule inferences, achieving an accuracy of approximately 75.95% in rule inference. In addition, an ablation study validates the effectiveness of the proposed approach in enhancing performance, confirming its potential as a reliable tool for early developmental screenings and clinical assessments. As such, the proposed approach could enhance clinical decision-making by providing essential interpretability for developmental tests.

{"title":"Rule training by score-based supervised contrastive learning for sketch explanation","authors":"Tae-Gyun Lee, Jang-Hee Yoo","doi":"10.1016/j.engappai.2025.110310","DOIUrl":"10.1016/j.engappai.2025.110310","url":null,"abstract":"<div><div>This paper presents a novel approach to explain scoring results of infant visual-motor integration sketches utilized in developmental tests by training predefined rules for each test item. To address the performance issues caused by limited data, we employ a pre-trained model that uses supervised contrastive learning based on item scores. To ensure effective training, a memory bank structure is proposed to accumulate diverse embeddings over multiple iterations and prevent the encoder that processes item information from being trained to prevent collapsing in the Siamese network. Experiments demonstrate that the proposed method improves performance in both score and rule inferences, achieving an accuracy of approximately 75.95% in rule inference. In addition, an ablation study validates the effectiveness of the proposed approach in enhancing performance, confirming its potential as a reliable tool for early developmental screenings and clinical assessments. As such, the proposed approach could enhance clinical decision-making by providing essential interpretability for developmental tests.</div></div>","PeriodicalId":50523,"journal":{"name":"Engineering Applications of Artificial Intelligence","volume":"147 ","pages":"Article 110310"},"PeriodicalIF":7.5,"publicationDate":"2025-02-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143489025","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}

引用次数: 0

A residual learning-based grey system model and its applications in Electricity Transformer’s Seasonal oil temperature forecasting

IF 7.5 2区计算机科学 Q1 AUTOMATION & CONTROL SYSTEMS

Engineering Applications of Artificial Intelligence

Pub Date : 2025-02-25 DOI: 10.1016/j.engappai.2025.110260

Yiwu Hao, Xin Ma, Lili Song, Yushu Xiang

Accurately predicting cross-regional electricity demand is crucial for efficient distribution management, but it remains challenging due to its complexity. Transformer oil temperature is a key indicator of operational status, and analyzing its seasonal variation is vital for addressing distribution issues. Grey models based on neural networks are effective for predicting nonlinear and small-scale datasets but are prone to overfitting. While residual networks help mitigate overfitting, their application to small-scale time series forecasting is still limited. To improve prediction accuracy for nonlinear and small-scale data, this study introduces residual learning into grey models, proposing a hybrid model. This model combines the feature-capturing ability of residual learning networks with the robustness of grey models, helping to reduce overfitting. The model is trained using the Adam algorithm, with parameters optimized by the Gridsearch algorithm. Performance is demonstrated using four seasonal datasets of transformer oil temperature. A comparison with 13 grey system models and 9 machine learning models shows that the proposed method outperforms the others. By calculating the percentage improvements of various metrics, the model demonstrates consistent performance gains. Sensitivity analysis reveals that the model’s performance is sensitive to the number of neurons and network depth, with higher values significantly improving accuracy and robustness. The results confirm the model’s effectiveness. This study fills the gap between neural grey models and residual networks, successfully applying the model to forecast the seasonal temperature trends of power transformers and providing a theoretical basis for addressing power distribution challenges.

{"title":"A residual learning-based grey system model and its applications in Electricity Transformer’s Seasonal oil temperature forecasting","authors":"Yiwu Hao, Xin Ma, Lili Song, Yushu Xiang","doi":"10.1016/j.engappai.2025.110260","DOIUrl":"10.1016/j.engappai.2025.110260","url":null,"abstract":"<div><div>Accurately predicting cross-regional electricity demand is crucial for efficient distribution management, but it remains challenging due to its complexity. Transformer oil temperature is a key indicator of operational status, and analyzing its seasonal variation is vital for addressing distribution issues. Grey models based on neural networks are effective for predicting nonlinear and small-scale datasets but are prone to overfitting. While residual networks help mitigate overfitting, their application to small-scale time series forecasting is still limited. To improve prediction accuracy for nonlinear and small-scale data, this study introduces residual learning into grey models, proposing a hybrid model. This model combines the feature-capturing ability of residual learning networks with the robustness of grey models, helping to reduce overfitting. The model is trained using the Adam algorithm, with parameters optimized by the Gridsearch algorithm. Performance is demonstrated using four seasonal datasets of transformer oil temperature. A comparison with 13 grey system models and 9 machine learning models shows that the proposed method outperforms the others. By calculating the percentage improvements of various metrics, the model demonstrates consistent performance gains. Sensitivity analysis reveals that the model’s performance is sensitive to the number of neurons and network depth, with higher values significantly improving accuracy and robustness. The results confirm the model’s effectiveness. This study fills the gap between neural grey models and residual networks, successfully applying the model to forecast the seasonal temperature trends of power transformers and providing a theoretical basis for addressing power distribution challenges.</div></div>","PeriodicalId":50523,"journal":{"name":"Engineering Applications of Artificial Intelligence","volume":"147 ","pages":"Article 110260"},"PeriodicalIF":7.5,"publicationDate":"2025-02-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143479002","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}

引用次数: 0

Evolution of Building Energy Management Systems for greater sustainability through explainable artificial intelligence models

IF 7.5 2区计算机科学 Q1 AUTOMATION & CONTROL SYSTEMS

Engineering Applications of Artificial Intelligence

Pub Date : 2025-02-25 DOI: 10.1016/j.engappai.2025.110324

Alfonso González-Briones , Javier Palomino-Sánchez , Zita Vale , Carlos Ramos , Juan M. Corchado

Predicting energy consumption is a task that allows energy supply companies to adapt to certain behaviours. The activities that companies can undertake include learning about the behaviour of their customers in order to adapt their tariffs to consumption or identifying the intervals in which there will be a higher demand for energy and to plan for the adaptation of supply chains. While predicting energy consumption is no longer a major challenge, and models with high accuracy rates have been developed, an clear understanding of energy consumption among users is still obscure. If the problem of explainability is resolved, companies will be able to better adapt their services by generating the exact amount of energy to be sold, which will also reduce its cost for customers. There is no single explanatory approach to learning models that works best. There are multiple paths to achieving explainability: data vs. model, directly interpretable vs. post hoc explanation, local vs. global, etc. This article reviews which explainable artificial intelligence algorithms are the most appropriate for a given use case, as multiple forms of explanation can lead to confusion in figuring out which algorithms are the most appropriate for a given use case. In our case study, a specific dataset, extracted from a two-year period in a shoe store, is used to review some of the main explainable artificial intelligence algorithms on machine learning models, capable of predicting energy consumption and subsequently providing explainability to the process.

{"title":"Evolution of Building Energy Management Systems for greater sustainability through explainable artificial intelligence models","authors":"Alfonso González-Briones , Javier Palomino-Sánchez , Zita Vale , Carlos Ramos , Juan M. Corchado","doi":"10.1016/j.engappai.2025.110324","DOIUrl":"10.1016/j.engappai.2025.110324","url":null,"abstract":"<div><div>Predicting energy consumption is a task that allows energy supply companies to adapt to certain behaviours. The activities that companies can undertake include learning about the behaviour of their customers in order to adapt their tariffs to consumption or identifying the intervals in which there will be a higher demand for energy and to plan for the adaptation of supply chains. While predicting energy consumption is no longer a major challenge, and models with high accuracy rates have been developed, an clear understanding of energy consumption among users is still obscure. If the problem of explainability is resolved, companies will be able to better adapt their services by generating the exact amount of energy to be sold, which will also reduce its cost for customers. There is no single explanatory approach to learning models that works best. There are multiple paths to achieving explainability: data vs. model, directly interpretable vs. post hoc explanation, local vs. global, etc. This article reviews which explainable artificial intelligence algorithms are the most appropriate for a given use case, as multiple forms of explanation can lead to confusion in figuring out which algorithms are the most appropriate for a given use case. In our case study, a specific dataset, extracted from a two-year period in a shoe store, is used to review some of the main explainable artificial intelligence algorithms on machine learning models, capable of predicting energy consumption and subsequently providing explainability to the process.</div></div>","PeriodicalId":50523,"journal":{"name":"Engineering Applications of Artificial Intelligence","volume":"147 ","pages":"Article 110324"},"PeriodicalIF":7.5,"publicationDate":"2025-02-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143478846","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}

引用次数: 0

Geometry-sensitive semantic modeling in visual and visual-language domains for image captioning

IF 7.5 2区计算机科学 Q1 AUTOMATION & CONTROL SYSTEMS

Engineering Applications of Artificial Intelligence

Pub Date : 2025-02-25 DOI: 10.1016/j.engappai.2025.110330

Wencai Zhu, Zetao Jiang, Yuting He

Transformer-based models with grid features as visual representations perform well in image captioning. However, the division and flattening operations increase the difficulty of capturing objects and their relationships via pure semantic modeling. Furthermore, the natural language generated by the current Transformer model still suffers from semantic overconcentration. In this paper, we aim to improve the attention modules in two ways to solve the above issues. We first propose a Geometry-Sensitive Self-Attention (GSSA) module, subdivide geometric signals in the visual domain into relative position and distance, and assist the semantic modeling process according to their unique characteristics. It compensates for the lack of objects and their relationships in the grid features. Then, we propose a Geometry-Sensitive Cross-Attention (GSCA) module, which perceives the source neighboring relationships between images and text in the visual-language domain from a geometric perspective and uses these relationships to adjust the semantic correspondences between the two dynamically. It spreads overly focused attention to surrounding grids to improve understanding of full image content during captioning. To prove our designs, we apply GSSA and GSCA to a standard Transformer to construct a novel Geometry-Sensitive Transformer Network (GSTNet), which conducts geometry-sensitive semantic modeling in visual and visual-language domains. Extensive experiments are conducted to verify the effectiveness of our proposal. The results show that our GSTNet achieves superior performance compared to many state-of-the-art image captioning models on the Microsoft Common Objects in Context (MSCOCO) dataset. Besides, the generalization of GSTNet is also verified on the Flickr30k dataset.

{"title":"Geometry-sensitive semantic modeling in visual and visual-language domains for image captioning","authors":"Wencai Zhu, Zetao Jiang, Yuting He","doi":"10.1016/j.engappai.2025.110330","DOIUrl":"10.1016/j.engappai.2025.110330","url":null,"abstract":"<div><div>Transformer-based models with grid features as visual representations perform well in image captioning. However, the division and flattening operations increase the difficulty of capturing objects and their relationships via pure semantic modeling. Furthermore, the natural language generated by the current Transformer model still suffers from semantic overconcentration. In this paper, we aim to improve the attention modules in two ways to solve the above issues. We first propose a Geometry-Sensitive Self-Attention (GSSA) module, subdivide geometric signals in the visual domain into relative position and distance, and assist the semantic modeling process according to their unique characteristics. It compensates for the lack of objects and their relationships in the grid features. Then, we propose a Geometry-Sensitive Cross-Attention (GSCA) module, which perceives the source neighboring relationships between images and text in the visual-language domain from a geometric perspective and uses these relationships to adjust the semantic correspondences between the two dynamically. It spreads overly focused attention to surrounding grids to improve understanding of full image content during captioning. To prove our designs, we apply GSSA and GSCA to a standard Transformer to construct a novel Geometry-Sensitive Transformer Network (GSTNet), which conducts geometry-sensitive semantic modeling in visual and visual-language domains. Extensive experiments are conducted to verify the effectiveness of our proposal. The results show that our GSTNet achieves superior performance compared to many state-of-the-art image captioning models on the Microsoft Common Objects in Context (MSCOCO) dataset. Besides, the generalization of GSTNet is also verified on the Flickr30k dataset.</div></div>","PeriodicalId":50523,"journal":{"name":"Engineering Applications of Artificial Intelligence","volume":"147 ","pages":"Article 110330"},"PeriodicalIF":7.5,"publicationDate":"2025-02-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143479004","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}

引用次数: 0

Graph Neural Networks for multivariate time-series forecasting via learning hierarchical spatiotemporal dependencies

IF 7.5 2区计算机科学 Q1 AUTOMATION & CONTROL SYSTEMS

Engineering Applications of Artificial Intelligence

Pub Date : 2025-02-25 DOI: 10.1016/j.engappai.2025.110304

Zhou Zhou , Ronisha Basker , Dit-Yan Yeung

Multivariate time-series forecasting is one of the essential tasks to draw insights from sequential data. Spatiotemporal Graph Neural Networks (STGNN) have attracted much attention in this field due to their capability to capture the underlying spatiotemporal dependencies. However, current STGNN solutions succumb to a higher degree of error in their predictions due to insufficient modelling of the dependencies and dynamics at different levels. In this paper, a Graph Neural Networks-based model is proposed for multivariate time-series forecasting via learning hierarchical spatiotemporal dependencies (HSDGNN). Specifically, variables are organised as nodes in a graph while each node serves as a subgraph consisting of the attributes of variables. Then two-level convolutions are designed on the hierarchical graph to model the spatial dependencies with different granularities. The changes in graph topologies are also encoded for strengthening dependency modelling across time and spatial dimensions. The proposed model is tested using real-world datasets from different domains, including transportation, electricity, and meteorology. The experimental results demonstrate that HSDGNN can outperform state-of-the-art baselines by up to 15.3% in terms of prediction accuracy, without compromising model scalability.

{"title":"Graph Neural Networks for multivariate time-series forecasting via learning hierarchical spatiotemporal dependencies","authors":"Zhou Zhou , Ronisha Basker , Dit-Yan Yeung","doi":"10.1016/j.engappai.2025.110304","DOIUrl":"10.1016/j.engappai.2025.110304","url":null,"abstract":"<div><div>Multivariate time-series forecasting is one of the essential tasks to draw insights from sequential data. Spatiotemporal Graph Neural Networks (STGNN) have attracted much attention in this field due to their capability to capture the underlying spatiotemporal dependencies. However, current STGNN solutions succumb to a higher degree of error in their predictions due to insufficient modelling of the dependencies and dynamics at different levels. In this paper, a Graph Neural Networks-based model is proposed for multivariate time-series forecasting via learning hierarchical spatiotemporal dependencies (HSDGNN). Specifically, variables are organised as nodes in a graph while each node serves as a subgraph consisting of the attributes of variables. Then two-level convolutions are designed on the hierarchical graph to model the spatial dependencies with different granularities. The changes in graph topologies are also encoded for strengthening dependency modelling across time and spatial dimensions. The proposed model is tested using real-world datasets from different domains, including transportation, electricity, and meteorology. The experimental results demonstrate that HSDGNN can outperform state-of-the-art baselines by up to 15.3% in terms of prediction accuracy, without compromising model scalability.</div></div>","PeriodicalId":50523,"journal":{"name":"Engineering Applications of Artificial Intelligence","volume":"147 ","pages":"Article 110304"},"PeriodicalIF":7.5,"publicationDate":"2025-02-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143489022","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}

引用次数: 0

MiFDeU: Multi-information fusion network based on dual-encoder for pelvic bones segmentation

IF 7.5 2区计算机科学 Q1 AUTOMATION & CONTROL SYSTEMS

Engineering Applications of Artificial Intelligence

Pub Date : 2025-02-25 DOI: 10.1016/j.engappai.2025.110230

Fujiao Ju , Yichu Wu , Mingjie Dong , Jingxin Zhao

The segmentation of bone fragments is crucial for preoperative planning and intraoperative navigation in reduction surgery. Recent advances in medical segmentation have predominantly focused on U-shaped frameworks that employ convolutional neural networks or transformer variants as the backbone. However, these frameworks, which rely on a single encoder, often struggle with integrating information from diverse features and processing irregular shapes in visual objects. Such limitations can reduce segmentation accuracy and impair generalization performance across different datasets. To address these issues, we introduce a multi-information fusion network based on dual-encoder for pelvic bones segmentation. In order to capture global contextual information and local features simultaneously, our model takes a light resnet and a graph neural network with swin-pool module as dual-encoder for effectively representing the global and local topologies. We construct a high-low multi-dimensional paired attention in the bottleneck for fusing spatial and channel information from different dimensions. Instead of using the traditional dice loss in the unet-like architecture, our model employs both topological loss and boundary loss to enhance the goal optimization. In the experiments, our model achieves a substantially lower dice similarity coefficient and comparable 95 % Hausdorff distance compared to other state-of-the-art. The experiments on across datasets verify the superiority and generalization of the proposed model.

{"title":"MiFDeU: Multi-information fusion network based on dual-encoder for pelvic bones segmentation","authors":"Fujiao Ju , Yichu Wu , Mingjie Dong , Jingxin Zhao","doi":"10.1016/j.engappai.2025.110230","DOIUrl":"10.1016/j.engappai.2025.110230","url":null,"abstract":"<div><div>The segmentation of bone fragments is crucial for preoperative planning and intraoperative navigation in reduction surgery. Recent advances in medical segmentation have predominantly focused on U-shaped frameworks that employ convolutional neural networks or transformer variants as the backbone. However, these frameworks, which rely on a single encoder, often struggle with integrating information from diverse features and processing irregular shapes in visual objects. Such limitations can reduce segmentation accuracy and impair generalization performance across different datasets. To address these issues, we introduce a multi-information fusion network based on dual-encoder for pelvic bones segmentation. In order to capture global contextual information and local features simultaneously, our model takes a light resnet and a graph neural network with swin-pool module as dual-encoder for effectively representing the global and local topologies. We construct a high-low multi-dimensional paired attention in the bottleneck for fusing spatial and channel information from different dimensions. Instead of using the traditional dice loss in the unet-like architecture, our model employs both topological loss and boundary loss to enhance the goal optimization. In the experiments, our model achieves a substantially lower dice similarity coefficient and comparable 95 % Hausdorff distance compared to other state-of-the-art. The experiments on across datasets verify the superiority and generalization of the proposed model.</div></div>","PeriodicalId":50523,"journal":{"name":"Engineering Applications of Artificial Intelligence","volume":"147 ","pages":"Article 110230"},"PeriodicalIF":7.5,"publicationDate":"2025-02-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143478847","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}

引用次数: 0