Pub Date : 2025-01-27DOI: 10.1109/ACCESS.2025.3534261
Rajesh Garapati;Manomita Chakraborty
In the era of digital technology, when material created by users is prevalent on online platforms, considerable difficulty is faced in analyzing large volumes of text in order to comprehend user emotions and forecast product ratings. The rapid rise in online reviews and comments necessitates the use of advanced tools to assess this data and extract valuable insights. This is considered crucial for the effectiveness of recommendation systems in many industries. This paper introduces the Review Text Granularity (RTG) Model, a new way to use the complex information in review texts to improve sentiment analysis and rating prediction. The RTG Model uses an advanced approach to scoring sentiments. It measures the strength of sentiments and gives a continuous sentiment score instead of a simple positive or negative label. This makes it different from other binary sentiment analysis methods. Multiple predictive modeling techniques are used, which makes it possible for this comprehensive sentiment analysis to greatly improve the accuracy of rating predictions. It has been shown by research that the depth of textual reviews is better captured and measured by the RTG Model, providing a more detailed and precise picture of user opinions. The RTG Model works really well, making rating predictions in recommendation systems more accurate and useful. A detailed study using a real-world dataset of IMDb movie reviews demonstrated this. The study emphasizes the benefits of utilizing intricate sentiment scores in addition to conventional rating data. The potential applicability of the RTG Model in several fields, including entertainment, e-commerce, and social media, is demonstrated, leading to enhanced and tailored user experiences.
{"title":"Enhancing Sentiment Analysis and Rating Prediction Using the Review Text Granularity (RTG) Model","authors":"Rajesh Garapati;Manomita Chakraborty","doi":"10.1109/ACCESS.2025.3534261","DOIUrl":"https://doi.org/10.1109/ACCESS.2025.3534261","url":null,"abstract":"In the era of digital technology, when material created by users is prevalent on online platforms, considerable difficulty is faced in analyzing large volumes of text in order to comprehend user emotions and forecast product ratings. The rapid rise in online reviews and comments necessitates the use of advanced tools to assess this data and extract valuable insights. This is considered crucial for the effectiveness of recommendation systems in many industries. This paper introduces the Review Text Granularity (RTG) Model, a new way to use the complex information in review texts to improve sentiment analysis and rating prediction. The RTG Model uses an advanced approach to scoring sentiments. It measures the strength of sentiments and gives a continuous sentiment score instead of a simple positive or negative label. This makes it different from other binary sentiment analysis methods. Multiple predictive modeling techniques are used, which makes it possible for this comprehensive sentiment analysis to greatly improve the accuracy of rating predictions. It has been shown by research that the depth of textual reviews is better captured and measured by the RTG Model, providing a more detailed and precise picture of user opinions. The RTG Model works really well, making rating predictions in recommendation systems more accurate and useful. A detailed study using a real-world dataset of IMDb movie reviews demonstrated this. The study emphasizes the benefits of utilizing intricate sentiment scores in addition to conventional rating data. The potential applicability of the RTG Model in several fields, including entertainment, e-commerce, and social media, is demonstrated, leading to enhanced and tailored user experiences.","PeriodicalId":13079,"journal":{"name":"IEEE Access","volume":"13 ","pages":"20071-20100"},"PeriodicalIF":3.4,"publicationDate":"2025-01-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10854458","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143105781","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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}
Pub Date : 2025-01-27DOI: 10.1109/ACCESS.2025.3534803
Andrew Prasetyo;I Ketut Eddy Purnama;Eko Mulyanto Yuniarno;Priyo Suprobo
Monitoring the health of infrastructure is critical to maintaining the integrity of concrete construction. Conventional crack detection methods that rely on visual inspection and image processing often produce inconsistent results. U-Net, an architecture often used in image processing, has limitations in crack segmentation, especially regarding the loss function. The main challenge is class imbalance, as cracks usually occupy a much smaller area compared to the background in concrete images. To address this, we developed an automated technique utilizing the U-Net framework with a novel loss function that we named DBCE Loss. Unlike the typical BCED we added the LogCoshDice function to get an improvement so that the resulting loss will be better and can improve the performance of the model. The performance of our method is rigorously evaluated by combining DeepCrack, GAPS, Crack500, and CrackForest datasets to illustrate that the model can detect cracks in various material conditions. DeepCrack represents cracks in concrete with minimal distress, while GAPS represents cracks in asphalt. Crack500 covers cracks with significant distress such as gravel, and CrackForest focuses on small cracks in concrete.The proposed U-Net model achieved Maximum accuracies between 98.58% and 99.22%, Maximum Dice coefficients from 88.27% to 98.03%, and Maximum F1-scores up to 97.11%. The proposed method is 7.69% ahead of the largest comparison method (DICE Loss) on Average IOU, while on Average Precission is leading 6.72% ahead of the largest comparison method (DICE). Then in terms of Recall our method is 8.84% superior to the largest comparison method (BCE). Sera in terms of Average F1-Score is 14.64% ahead of the largest comparison method (DICE). These results show that our method has surpassed conventional loss methods such as BCE, DICE and BCED. This research can facilitate a more reliable infrastructure health monitoring process and help with crack detection in the future.
{"title":"Improving Crack Detection Precision of Concrete Structures Using U-Net Architecture and Novel DBCE Loss Function","authors":"Andrew Prasetyo;I Ketut Eddy Purnama;Eko Mulyanto Yuniarno;Priyo Suprobo","doi":"10.1109/ACCESS.2025.3534803","DOIUrl":"https://doi.org/10.1109/ACCESS.2025.3534803","url":null,"abstract":"Monitoring the health of infrastructure is critical to maintaining the integrity of concrete construction. Conventional crack detection methods that rely on visual inspection and image processing often produce inconsistent results. U-Net, an architecture often used in image processing, has limitations in crack segmentation, especially regarding the loss function. The main challenge is class imbalance, as cracks usually occupy a much smaller area compared to the background in concrete images. To address this, we developed an automated technique utilizing the U-Net framework with a novel loss function that we named DBCE Loss. Unlike the typical BCED we added the LogCoshDice function to get an improvement so that the resulting loss will be better and can improve the performance of the model. The performance of our method is rigorously evaluated by combining DeepCrack, GAPS, Crack500, and CrackForest datasets to illustrate that the model can detect cracks in various material conditions. DeepCrack represents cracks in concrete with minimal distress, while GAPS represents cracks in asphalt. Crack500 covers cracks with significant distress such as gravel, and CrackForest focuses on small cracks in concrete.The proposed U-Net model achieved Maximum accuracies between 98.58% and 99.22%, Maximum Dice coefficients from 88.27% to 98.03%, and Maximum F1-scores up to 97.11%. The proposed method is 7.69% ahead of the largest comparison method (DICE Loss) on Average IOU, while on Average Precission is leading 6.72% ahead of the largest comparison method (DICE). Then in terms of Recall our method is 8.84% superior to the largest comparison method (BCE). Sera in terms of Average F1-Score is 14.64% ahead of the largest comparison method (DICE). These results show that our method has surpassed conventional loss methods such as BCE, DICE and BCED. This research can facilitate a more reliable infrastructure health monitoring process and help with crack detection in the future.","PeriodicalId":13079,"journal":{"name":"IEEE Access","volume":"13 ","pages":"20903-20922"},"PeriodicalIF":3.4,"publicationDate":"2025-01-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10855421","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143105909","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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}
Pub Date : 2025-01-27DOI: 10.1109/ACCESS.2025.3534782
Xin Geng;Xiao Han;Xianghong Cao;Yixuan Su;Dongxue Shu
Regarding the current problems of false alarms and missed detections in fire detection, we propose a high-precision fire detection algorithm, YOLOV9-CBM (C3-SE, BiFPN, MPDIoU), by optimizing YOLOV9. Firstly, to tackle the shortage of both quality and quantity in the existing fire datasets, we collected 2,000 fire and smoke images to establish a dataset named CBM-Fire. Secondly, the RepNCSPELAN4 module of the YOLOv9 backbone was replaced with the C3 module containing SE Attention to improve detection efficiency while guaranteeing accuracy. Besides, we transformed the multi-scale fusion network PANet in the baseline algorithm into a bidirectional feature network pyramid BiFPN to facilitate the bidirectional flow of features, enabling the algorithm to fuse information at different scales more effectively. Finally, instead of CIoU losses, we adopted MPDIoU losses in bounding box regression, which improved the accuracy of model regression and classification. Experimental results indicate that compared with YOLOV9, the recall rate of YOLOV9-CBM has increased by 7.6% and the mAP has risen by 3.8%. The revised model demonstrates good generalization performance and robustness. Code and dataset are at https://github.com/GengHan-123/yolov9-cbm.git.
{"title":"YOLOV9-CBM: An Improved Fire Detection Algorithm Based on YOLOV9","authors":"Xin Geng;Xiao Han;Xianghong Cao;Yixuan Su;Dongxue Shu","doi":"10.1109/ACCESS.2025.3534782","DOIUrl":"https://doi.org/10.1109/ACCESS.2025.3534782","url":null,"abstract":"Regarding the current problems of false alarms and missed detections in fire detection, we propose a high-precision fire detection algorithm, YOLOV9-CBM (C3-SE, BiFPN, MPDIoU), by optimizing YOLOV9. Firstly, to tackle the shortage of both quality and quantity in the existing fire datasets, we collected 2,000 fire and smoke images to establish a dataset named CBM-Fire. Secondly, the RepNCSPELAN4 module of the YOLOv9 backbone was replaced with the C3 module containing SE Attention to improve detection efficiency while guaranteeing accuracy. Besides, we transformed the multi-scale fusion network PANet in the baseline algorithm into a bidirectional feature network pyramid BiFPN to facilitate the bidirectional flow of features, enabling the algorithm to fuse information at different scales more effectively. Finally, instead of CIoU losses, we adopted MPDIoU losses in bounding box regression, which improved the accuracy of model regression and classification. Experimental results indicate that compared with YOLOV9, the recall rate of YOLOV9-CBM has increased by 7.6% and the mAP has risen by 3.8%. The revised model demonstrates good generalization performance and robustness. Code and dataset are at <uri>https://github.com/GengHan-123/yolov9-cbm.git</uri>.","PeriodicalId":13079,"journal":{"name":"IEEE Access","volume":"13 ","pages":"19612-19623"},"PeriodicalIF":3.4,"publicationDate":"2025-01-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10854439","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143105501","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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}
Pub Date : 2025-01-27DOI: 10.1109/ACCESS.2025.3535314
Antonio V. A. Lundgren;Byron L. D. Bezerra;Carmelo J. A. Bastos-Filho
We introduce VisualSAF, a novel Visual Semantic Analysis Framework designed to enhance the understanding of contextual characteristics in Visual Scene Analysis (VSA) tasks. The framework leverages semantic variables extracted using machine learning algorithms to provide additional high-level information, augmenting the capabilities of the primary task model. Comprising three main components – the General DL Model, Semantic Variables, and Output Branches – VisualSAF offers a modular and adaptable approach to addressing diverse VSA tasks. The General DL Model processes input images, extracting high-level features through a backbone network and detecting regions of interest. Semantic Variables are then extracted from these regions, incorporating a wide range of contextual information tailored to specific scenarios. Finally, the Output Branch integrates semantic variables and detections, generating high-level task information while allowing for flexible weighting of inputs to optimize task performance. The framework is demonstrated through experiments on the HOD Dataset, showcasing improvements in mean average precision and mean average recall compared to baseline models; the improvements are 0.05 in both mAP and 0.01 in mAR compared to the baseline. Future research directions include exploring multiple semantic variables, developing more complex output heads, and investigating the framework’s performance across context-shifting datasets.
{"title":"VisualSAF-A Novel Framework for Visual Semantic Analysis Tasks","authors":"Antonio V. A. Lundgren;Byron L. D. Bezerra;Carmelo J. A. Bastos-Filho","doi":"10.1109/ACCESS.2025.3535314","DOIUrl":"https://doi.org/10.1109/ACCESS.2025.3535314","url":null,"abstract":"We introduce VisualSAF, a novel Visual Semantic Analysis Framework designed to enhance the understanding of contextual characteristics in Visual Scene Analysis (VSA) tasks. The framework leverages semantic variables extracted using machine learning algorithms to provide additional high-level information, augmenting the capabilities of the primary task model. Comprising three main components – the General DL Model, Semantic Variables, and Output Branches – VisualSAF offers a modular and adaptable approach to addressing diverse VSA tasks. The General DL Model processes input images, extracting high-level features through a backbone network and detecting regions of interest. Semantic Variables are then extracted from these regions, incorporating a wide range of contextual information tailored to specific scenarios. Finally, the Output Branch integrates semantic variables and detections, generating high-level task information while allowing for flexible weighting of inputs to optimize task performance. The framework is demonstrated through experiments on the HOD Dataset, showcasing improvements in mean average precision and mean average recall compared to baseline models; the improvements are 0.05 in both mAP and 0.01 in mAR compared to the baseline. Future research directions include exploring multiple semantic variables, developing more complex output heads, and investigating the framework’s performance across context-shifting datasets.","PeriodicalId":13079,"journal":{"name":"IEEE Access","volume":"13 ","pages":"21052-21063"},"PeriodicalIF":3.4,"publicationDate":"2025-01-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10855394","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143105734","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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}
Pub Date : 2025-01-27DOI: 10.1109/ACCESS.2025.3534630
Monia Digra;Upma Jain;Ram Kishan Dewangan;Himanshu Suyal
This paper proposes a hybrid approach for multiple Unmanned Aerial Vehicle navigation. This is an NP-hard problem since the robots must find the optimal safe path without colliding with other robots and obstacles in a three-dimensional search space. The proposed approach enhances the exploration capabilities of the whale optimization algorithm. Then, it hybridises this improved whale optimization algorithm with the sine cosine algorithm to improve the overall exploitation capabilities. The efficiency of the proposed hybrid approach is compared with other meta-heuristic algorithms for multi-UAV navigation. Results obtained through simulation ensure the validity of the proposed approach.
{"title":"UAV-NavS: Three-Dimensional Navigation System of Multiple Unmanned Aerial Vehicles Using Hybrid Optimization Algorithm","authors":"Monia Digra;Upma Jain;Ram Kishan Dewangan;Himanshu Suyal","doi":"10.1109/ACCESS.2025.3534630","DOIUrl":"https://doi.org/10.1109/ACCESS.2025.3534630","url":null,"abstract":"This paper proposes a hybrid approach for multiple Unmanned Aerial Vehicle navigation. This is an NP-hard problem since the robots must find the optimal safe path without colliding with other robots and obstacles in a three-dimensional search space. The proposed approach enhances the exploration capabilities of the whale optimization algorithm. Then, it hybridises this improved whale optimization algorithm with the sine cosine algorithm to improve the overall exploitation capabilities. The efficiency of the proposed hybrid approach is compared with other meta-heuristic algorithms for multi-UAV navigation. Results obtained through simulation ensure the validity of the proposed approach.","PeriodicalId":13079,"journal":{"name":"IEEE Access","volume":"13 ","pages":"20247-20259"},"PeriodicalIF":3.4,"publicationDate":"2025-01-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10854430","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143106014","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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}
The multi-day intermodal trip planning problem (MITPP) is an optimization problem that seeks to create the optimal route to visit Point-of-Interest (POI) and hotels over days. This problem involves coordinating intermodal transportation, such as walking, public transportation, to create a well-crafted travel itinerary. Quantum annealers have recently been explored as a powerful tool for solving combinatorial optimization problems by converting the problems into Quadratic Unconstrained Binary Optimization (QUBO). However, current quantum annealers have a small QUBO input size so that they cannot directly solve large-scale MITPPs. In this paper, we address this issue by extracting a subQUBO from the original large QUBO based on variable (spin) deviations and randomness. Then, we iteratively solve the subQUBOs by the quantum annealer and update the (quasi-)optimal solution. As the obtained (quasi-)optimal solution may violate constraints, we apply the correction processing till all constraints are satisfied. According to the experiment results using a real quantum annealer, our proposed method obtained high-quality solutions for large-scale MITPPs in the Tokyo area, and compared to the full QUBO method, we achieve a maximum spin reduction of 98.9%. Especially, compared to the method by a conventional computer and two conventional subQUBO methods, POI satisfaction is improved by 10.2%, and travel costs are improved by 23.2% respectively.
{"title":"Hybrid subQUBO Annealing With a Correction Process for Multi-Day Intermodal Trip Planning","authors":"Tatsuya Noguchi;Keisuke Fukada;Siya Bao;Nozomu Togawa","doi":"10.1109/ACCESS.2025.3534529","DOIUrl":"https://doi.org/10.1109/ACCESS.2025.3534529","url":null,"abstract":"The multi-day intermodal trip planning problem (MITPP) is an optimization problem that seeks to create the optimal route to visit Point-of-Interest (POI) and hotels over days. This problem involves coordinating intermodal transportation, such as walking, public transportation, to create a well-crafted travel itinerary. Quantum annealers have recently been explored as a powerful tool for solving combinatorial optimization problems by converting the problems into Quadratic Unconstrained Binary Optimization (QUBO). However, current quantum annealers have a small QUBO input size so that they cannot directly solve large-scale MITPPs. In this paper, we address this issue by extracting a subQUBO from the original large QUBO based on variable (spin) deviations and randomness. Then, we iteratively solve the subQUBOs by the quantum annealer and update the (quasi-)optimal solution. As the obtained (quasi-)optimal solution may violate constraints, we apply the correction processing till all constraints are satisfied. According to the experiment results using a real quantum annealer, our proposed method obtained high-quality solutions for large-scale MITPPs in the Tokyo area, and compared to the full QUBO method, we achieve a maximum spin reduction of 98.9%. Especially, compared to the method by a conventional computer and two conventional subQUBO methods, POI satisfaction is improved by 10.2%, and travel costs are improved by 23.2% respectively.","PeriodicalId":13079,"journal":{"name":"IEEE Access","volume":"13 ","pages":"19716-19727"},"PeriodicalIF":3.4,"publicationDate":"2025-01-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10854423","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143105854","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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}
Integrity monitoring is a key criterion for achieving robust and safe navigation systems. This work explores two integrity frameworks: the classical methods and their respective evolution towards the road vehicle urban scenario, and the artificial intelligence-based methods, where the monitoring process is accomplished by data analysis and learning techniques. In most cases, machine learning outperforms traditional models, which are often observed under controlled, non-real-time conditions, by employing simple algorithms that may have limited success in real-world applications. An overview is provided on how these algorithms have been used, including a comparison of their characteristics and performances, offering insights into how they can evolve and possible future directions to achieve more reliable solutions.
{"title":"Smart GNSS Integrity Monitoring for Road Vehicles: An Overview of AI Methods","authors":"Inês Viveiros;Hélder Silva;Yuri Andrade;Cristiano Pendão","doi":"10.1109/ACCESS.2025.3534659","DOIUrl":"https://doi.org/10.1109/ACCESS.2025.3534659","url":null,"abstract":"Integrity monitoring is a key criterion for achieving robust and safe navigation systems. This work explores two integrity frameworks: the classical methods and their respective evolution towards the road vehicle urban scenario, and the artificial intelligence-based methods, where the monitoring process is accomplished by data analysis and learning techniques. In most cases, machine learning outperforms traditional models, which are often observed under controlled, non-real-time conditions, by employing simple algorithms that may have limited success in real-world applications. An overview is provided on how these algorithms have been used, including a comparison of their characteristics and performances, offering insights into how they can evolve and possible future directions to achieve more reliable solutions.","PeriodicalId":13079,"journal":{"name":"IEEE Access","volume":"13 ","pages":"20278-20296"},"PeriodicalIF":3.4,"publicationDate":"2025-01-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10854211","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143105970","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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}
Pub Date : 2025-01-27DOI: 10.1109/ACCESS.2025.3534880
Muhammad Asif;Talha Ahmed Khan;Wang-Cheol Song
Intent-Based Networking (IBN) simplifies network management by enabling users to express high-level intents in natural language, but existing approaches often fail to ensure alignment with network policies, leading to misconfigurations. Moreover, many methods lack robust validation mechanisms, reducing their reliability in dynamic environments. This research addresses these gaps by evaluating advanced Large Language Models (LLMs) such as BERT-base uncased (BERT-bu), GPT2, LLaMA3, Claude2 and small deep learning model BiLSTM with attention for translating intents and detecting contradictions. Using a curated dataset of 10,000 intent pairs, the proposed hybrid framework integrates a K-Nearest Neighbors (KNN) classifier to validate translations and recalibrate erroneous outputs. Experimental results demonstrate up to 5% higher accuracy (88%) and F1 scores compared to existing methods, ensuring precise intent translation and reliable network orchestration. This approach significantly enhances scalability and policy compliance in automated network environments.
{"title":"Evaluating Large Language Models for Optimized Intent Translation and Contradiction Detection Using KNN in IBN","authors":"Muhammad Asif;Talha Ahmed Khan;Wang-Cheol Song","doi":"10.1109/ACCESS.2025.3534880","DOIUrl":"https://doi.org/10.1109/ACCESS.2025.3534880","url":null,"abstract":"Intent-Based Networking (IBN) simplifies network management by enabling users to express high-level intents in natural language, but existing approaches often fail to ensure alignment with network policies, leading to misconfigurations. Moreover, many methods lack robust validation mechanisms, reducing their reliability in dynamic environments. This research addresses these gaps by evaluating advanced Large Language Models (LLMs) such as BERT-base uncased (BERT-bu), GPT2, LLaMA3, Claude2 and small deep learning model BiLSTM with attention for translating intents and detecting contradictions. Using a curated dataset of 10,000 intent pairs, the proposed hybrid framework integrates a K-Nearest Neighbors (KNN) classifier to validate translations and recalibrate erroneous outputs. Experimental results demonstrate up to 5% higher accuracy (88%) and F1 scores compared to existing methods, ensuring precise intent translation and reliable network orchestration. This approach significantly enhances scalability and policy compliance in automated network environments.","PeriodicalId":13079,"journal":{"name":"IEEE Access","volume":"13 ","pages":"20316-20327"},"PeriodicalIF":3.4,"publicationDate":"2025-01-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10855447","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143106098","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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}
Pub Date : 2025-01-27DOI: 10.1109/ACCESS.2025.3534783
Yuta Abekura;Takanori Hara;Satoshi Suyama;Satoshi Nagata;Kenichi Higuchi
In this paper, we propose two peak-to-average power ratio (PAPR) reduction methods that apply adaptive in-band interference restriction based on singular value differences between spatial channels in uplink eigenmode massive multiple-input multiple-output (MIMO) transmission. In uplink transmission utilizing high-frequency bands, reducing PAPR is crucial to suppress non-linear distortion and ensure sufficient transmission distance, given the stringent amplification requirements of power amplifiers. Therefore, the proposed methods direct the PAPR reduction signals caused by clipping and filtering (CF) only to spatial channels with relatively small singular values, thereby avoiding in-band interference in spatial channels with larger singular values, which are more susceptible to such interference compared to the case without PAPR reduction. After that, the two proposed methods restrict PAPR reduction signals for all or a subset of subcarriers based on the transmission quality requirements. Both interference restrictions allow for more effective PAPR reduction while tolerating interference in the data streams. Computer simulations demonstrate that the transmission quality of the proposed methods is improved by up to approximately 30% compared to the case without PAPR reduction. Moreover, a comparative evaluation of the two proposed methods demonstrates that the interference restriction across all subcarriers effectively mitigates in-band interference caused by the resulting PAPR reduction signals while reducing the PAPR.
{"title":"Adaptive Interference Restriction Among Spatial Streams Based on Difference in Singular Values for PAPR Reduction in Uplink Eigenmode Massive MIMO Transmission","authors":"Yuta Abekura;Takanori Hara;Satoshi Suyama;Satoshi Nagata;Kenichi Higuchi","doi":"10.1109/ACCESS.2025.3534783","DOIUrl":"https://doi.org/10.1109/ACCESS.2025.3534783","url":null,"abstract":"In this paper, we propose two peak-to-average power ratio (PAPR) reduction methods that apply adaptive in-band interference restriction based on singular value differences between spatial channels in uplink eigenmode massive multiple-input multiple-output (MIMO) transmission. In uplink transmission utilizing high-frequency bands, reducing PAPR is crucial to suppress non-linear distortion and ensure sufficient transmission distance, given the stringent amplification requirements of power amplifiers. Therefore, the proposed methods direct the PAPR reduction signals caused by clipping and filtering (CF) only to spatial channels with relatively small singular values, thereby avoiding in-band interference in spatial channels with larger singular values, which are more susceptible to such interference compared to the case without PAPR reduction. After that, the two proposed methods restrict PAPR reduction signals for all or a subset of subcarriers based on the transmission quality requirements. Both interference restrictions allow for more effective PAPR reduction while tolerating interference in the data streams. Computer simulations demonstrate that the transmission quality of the proposed methods is improved by up to approximately 30% compared to the case without PAPR reduction. Moreover, a comparative evaluation of the two proposed methods demonstrates that the interference restriction across all subcarriers effectively mitigates in-band interference caused by the resulting PAPR reduction signals while reducing the PAPR.","PeriodicalId":13079,"journal":{"name":"IEEE Access","volume":"13 ","pages":"21520-21532"},"PeriodicalIF":3.4,"publicationDate":"2025-01-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10855449","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143105688","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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}
Pub Date : 2025-01-27DOI: 10.1109/ACCESS.2025.3534847
Andrius Čeponis;Dalius Mažeika;Regimantas Bareikis
Article introduces a novel 3-DOF piezoelectric actuator designed to achieve the high-precision angular motion of a spherical rotor and an optical mirror around three axes. The actuator consists of three interlinked piezoelectric bimorph plates arranged in a low-profile triangular structure. Each piezoelectric plate contains a cylindrical contact that is used to transfer the vibrations to induce the angular motion of the rotor. The actuator has a low profile and is mounted on a printed circuit board (PCB), enhancing its structural integrity. The actuator occupies a footprint of 986 mm2 and, including the rotor, weighs 35.5 g. The actuator operation is based on the inertial stick-slip principle, using the first and second out-of-plane bending modes of the bimorph plates. Simultaneous excitation of the second vibration mode of all three bimorph plates produces angular motion of the rotor around the vertical axis. In contrast, the excitation of the first vibration mode of the individual plate enables angular motion around one of the horizontal axes. Numerical analysis identified the vibration modes, resonant frequencies, and mechanical and electromechanical characteristics of the actuator, leading to design optimization of clamping beams and contact locations on the bimorph plates. Experimental measurements revealed that the maximum rotation speeds are 363.7 RPM around the vertical axis and 129.1 RPM around the horizontal axis. Maximum torques of 218.2 mN/m and 159.7 mN/m were achieved around the vertical and horizontal axes, respectively, when the actuator is driven at 200 V$_{mathrm {p-p}}$ . The actuator demonstrated angular resolutions of 2.47 mrad and 1.033 mrad for the vertical and horizontal axes, respectively.
{"title":"3-DOF Inertial Piezoelectric Actuator for Angular Positioning of the Optical Mirror","authors":"Andrius Čeponis;Dalius Mažeika;Regimantas Bareikis","doi":"10.1109/ACCESS.2025.3534847","DOIUrl":"https://doi.org/10.1109/ACCESS.2025.3534847","url":null,"abstract":"Article introduces a novel 3-DOF piezoelectric actuator designed to achieve the high-precision angular motion of a spherical rotor and an optical mirror around three axes. The actuator consists of three interlinked piezoelectric bimorph plates arranged in a low-profile triangular structure. Each piezoelectric plate contains a cylindrical contact that is used to transfer the vibrations to induce the angular motion of the rotor. The actuator has a low profile and is mounted on a printed circuit board (PCB), enhancing its structural integrity. The actuator occupies a footprint of 986 mm2 and, including the rotor, weighs 35.5 g. The actuator operation is based on the inertial stick-slip principle, using the first and second out-of-plane bending modes of the bimorph plates. Simultaneous excitation of the second vibration mode of all three bimorph plates produces angular motion of the rotor around the vertical axis. In contrast, the excitation of the first vibration mode of the individual plate enables angular motion around one of the horizontal axes. Numerical analysis identified the vibration modes, resonant frequencies, and mechanical and electromechanical characteristics of the actuator, leading to design optimization of clamping beams and contact locations on the bimorph plates. Experimental measurements revealed that the maximum rotation speeds are 363.7 RPM around the vertical axis and 129.1 RPM around the horizontal axis. Maximum torques of 218.2 mN/m and 159.7 mN/m were achieved around the vertical and horizontal axes, respectively, when the actuator is driven at 200 V<inline-formula> <tex-math>$_{mathrm {p-p}}$ </tex-math></inline-formula>. The actuator demonstrated angular resolutions of 2.47 mrad and 1.033 mrad for the vertical and horizontal axes, respectively.","PeriodicalId":13079,"journal":{"name":"IEEE Access","volume":"13 ","pages":"20025-20039"},"PeriodicalIF":3.4,"publicationDate":"2025-01-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10855440","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143105768","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：