Lithium-ion batteries are widely used in industrial and domestic applications because of their high energy ratio and low self-discharge rate. It is important to accurately predict the State of Health (SoH) of lithium-ion batteries as they degrade during use, which can lead to serious safety hazards. We propose a support-based neural network ensemble method, which incorporates the prediction results of several basic neural network models. First, a set of better initial integration weights is calculated and the initial integration result is obtained, then the support degree between this result and the prediction result of each basic neural network is calculated, and the final integration weights are calculated by the weight iterative update ensemble algorithm and the integration prediction result of lithium-ion batteries SoH is obtained. This method avoids the risk of the "majority principle" which does not guarantee that most models perform better, and removes the constraint of positive integration weights, which can further reduce the adverse effects of poorly performing models on the integration results. We demonstrate the effectiveness of the proposed ensemble method for the lithium-ion batteries SoH prediction problem through a 5-fold cross-validation experiment on two datasets.
{"title":"Support-based Neural Network Ensemble Method for Predicting the SoH of Lithium-ion Battery","authors":"Hengshan Zhang, Jiaxuan Xu, Di Wu, Yun Wang","doi":"10.1145/3573942.3573958","DOIUrl":"https://doi.org/10.1145/3573942.3573958","url":null,"abstract":"Lithium-ion batteries are widely used in industrial and domestic applications because of their high energy ratio and low self-discharge rate. It is important to accurately predict the State of Health (SoH) of lithium-ion batteries as they degrade during use, which can lead to serious safety hazards. We propose a support-based neural network ensemble method, which incorporates the prediction results of several basic neural network models. First, a set of better initial integration weights is calculated and the initial integration result is obtained, then the support degree between this result and the prediction result of each basic neural network is calculated, and the final integration weights are calculated by the weight iterative update ensemble algorithm and the integration prediction result of lithium-ion batteries SoH is obtained. This method avoids the risk of the \"majority principle\" which does not guarantee that most models perform better, and removes the constraint of positive integration weights, which can further reduce the adverse effects of poorly performing models on the integration results. We demonstrate the effectiveness of the proposed ensemble method for the lithium-ion batteries SoH prediction problem through a 5-fold cross-validation experiment on two datasets.","PeriodicalId":103293,"journal":{"name":"Proceedings of the 2022 5th International Conference on Artificial Intelligence and Pattern Recognition","volume":"122 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122496991","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In order to enable the positioning system to complete real-time positioning in indoor and outdoor environments, an Ultra-wide band/Beidou indoor and outdoor seamless positioning algorithm based on fuzzy neural network and weighted cost function (FNN-WCF) is proposed. Through the analysis of multi-source data, the influencing factors needed in the network handover are determined, and the FNN-WCF handover model is constructed to realize seamless handover between different positioning subsystems. Perform operations such as fuzzing and defuzzifying each parameter to calculate the function value of different subsystems, and then compare the size of the function value to choose whether to switch. Experimental results show that the model can effectively realize the vertical switching between different positioning subsystems, and the positioning accuracy is less than 0.25m.
{"title":"Ultra-wideband/BDS Indoor and Outdoor Seamless Positioning Algorithm Based on FNN-WCF","authors":"Baojun Zhang, Haitao Zhan, Yinglong Hou, X. Chen","doi":"10.1145/3573942.3574001","DOIUrl":"https://doi.org/10.1145/3573942.3574001","url":null,"abstract":"In order to enable the positioning system to complete real-time positioning in indoor and outdoor environments, an Ultra-wide band/Beidou indoor and outdoor seamless positioning algorithm based on fuzzy neural network and weighted cost function (FNN-WCF) is proposed. Through the analysis of multi-source data, the influencing factors needed in the network handover are determined, and the FNN-WCF handover model is constructed to realize seamless handover between different positioning subsystems. Perform operations such as fuzzing and defuzzifying each parameter to calculate the function value of different subsystems, and then compare the size of the function value to choose whether to switch. Experimental results show that the model can effectively realize the vertical switching between different positioning subsystems, and the positioning accuracy is less than 0.25m.","PeriodicalId":103293,"journal":{"name":"Proceedings of the 2022 5th International Conference on Artificial Intelligence and Pattern Recognition","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130370125","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Aspect-level sentiment analysis is highly dependent on local context. However, most models are overly concerned with global context and external semantic knowledge. This approach increases the training time of the models. We propose the LGLFF (Lightweight Global and Local Feature Fusion) model. Firstly, we introduce a Distilroberta pretrained model in the LGLFF to encode the global context. Secondly, we use the SRU++ (Simple Recurrent Unit) network to extract global features. Then we adjust the SRD (Semantic-Relative Distance) threshold size by different datasets, and use SRD to mask the global context to get the local context. Finally, we use the multi-head attention mechanism to learn the global and local context features. We do some experiments on three datasets: Twitter, Laptop, and Restaurant. The results show that our model performs better than other benchmark models.
方面级情感分析高度依赖于本地上下文。然而，大多数模型都过于关注全局上下文和外部语义知识。这种方法增加了模型的训练时间。我们提出了LGLFF (Lightweight Global and Local Feature Fusion)模型。首先，我们在LGLFF中引入一个蒸馏roberta预训练模型来对全局上下文进行编码。其次，我们使用SRU++(简单循环单元)网络提取全局特征。然后根据不同的数据集调整SRD (Semantic-Relative Distance，语义相对距离)阈值大小，利用SRD掩盖全局上下文，得到局部上下文。最后，我们使用多头注意机制来学习全局和局部上下文特征。我们在三个数据集上做了一些实验:Twitter, Laptop和Restaurant。结果表明，该模型的性能优于其他基准模型。
{"title":"LGLFF: A Lightweight Aspect-Level Sentiment Analysis Model Based on Local-Global Features","authors":"Hao Liang, Xiaopeng Cao, Kaili Wang","doi":"10.1145/3573942.3573967","DOIUrl":"https://doi.org/10.1145/3573942.3573967","url":null,"abstract":"Aspect-level sentiment analysis is highly dependent on local context. However, most models are overly concerned with global context and external semantic knowledge. This approach increases the training time of the models. We propose the LGLFF (Lightweight Global and Local Feature Fusion) model. Firstly, we introduce a Distilroberta pretrained model in the LGLFF to encode the global context. Secondly, we use the SRU++ (Simple Recurrent Unit) network to extract global features. Then we adjust the SRD (Semantic-Relative Distance) threshold size by different datasets, and use SRD to mask the global context to get the local context. Finally, we use the multi-head attention mechanism to learn the global and local context features. We do some experiments on three datasets: Twitter, Laptop, and Restaurant. The results show that our model performs better than other benchmark models.","PeriodicalId":103293,"journal":{"name":"Proceedings of the 2022 5th International Conference on Artificial Intelligence and Pattern Recognition","volume":"183 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124639609","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
At present, wireless sensor networks are more and more favored by experts and scholars, and become a research hotspot in the field of sensing. Sensor networks are mainly used in environmental monitoring, wildlife detection and so on. When a sensor node is in a fast-moving environment, the node needs to discover its neighbors as quickly as possible. Therefore, neighbor discovery has attracted the attention of researchers. Neighbor discovery is an indispensable process in wireless sensor networks. Most of the current neighbor discovery designs are based on paired discovery and a fixed duty cycle. Only when two nodes wake up at the same time can they discover each other. This is completely passive neighbor discovery, and the network discovery delay is too large. And the nodes in the network are constantly moving. This is a challenging problem to reduce the discovery delay. This paper proposes a neighbor discovery algorithm (GDA, in short) that dynamically adjusts the wake-up time of nodes based on group spatial characteristics. At the same time, in order to effectively balance the relationship between energy consumption and discovery delay, a neighbor discovery algorithm that can selectively recommend method of neighbor nodes. This method can recommend suitable neighbor nodes and improve the early detection time. This paper elaborates the network model and algorithm implementation in detail. A large number of simulation results show that the algorithm has achieved good results in reducing discovery delay and network energy consumption.
{"title":"A Group-Based Dynamic Neighbor Discovery Algorithm in Mobile Sensor Networks","authors":"Shuai Li, Dongming Xu","doi":"10.1145/3573942.3573976","DOIUrl":"https://doi.org/10.1145/3573942.3573976","url":null,"abstract":"At present, wireless sensor networks are more and more favored by experts and scholars, and become a research hotspot in the field of sensing. Sensor networks are mainly used in environmental monitoring, wildlife detection and so on. When a sensor node is in a fast-moving environment, the node needs to discover its neighbors as quickly as possible. Therefore, neighbor discovery has attracted the attention of researchers. Neighbor discovery is an indispensable process in wireless sensor networks. Most of the current neighbor discovery designs are based on paired discovery and a fixed duty cycle. Only when two nodes wake up at the same time can they discover each other. This is completely passive neighbor discovery, and the network discovery delay is too large. And the nodes in the network are constantly moving. This is a challenging problem to reduce the discovery delay. This paper proposes a neighbor discovery algorithm (GDA, in short) that dynamically adjusts the wake-up time of nodes based on group spatial characteristics. At the same time, in order to effectively balance the relationship between energy consumption and discovery delay, a neighbor discovery algorithm that can selectively recommend method of neighbor nodes. This method can recommend suitable neighbor nodes and improve the early detection time. This paper elaborates the network model and algorithm implementation in detail. A large number of simulation results show that the algorithm has achieved good results in reducing discovery delay and network energy consumption.","PeriodicalId":103293,"journal":{"name":"Proceedings of the 2022 5th International Conference on Artificial Intelligence and Pattern Recognition","volume":"46 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116490730","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
As a new generation of automotive bus, the Flex Ray Alliance includes the largest automotive industry and the most influential. The Flex Ray bus has a very wide range of applications [1], which can promote the development of future automotive electronic systems. The MFR4310 stand-alone FlexRay controller enables easy integration of FlexRay into MCU-based applications and complies with the FlexRay Alliance specification. Driven by the wave of chip localization, in order to solve the problem of insufficient self-sufficiency of China's automotive-grade chips, a communication controller chip based on FlexRay bus was designed, which is fully compatible with MFR4310. In this paper, the implementation of each module of the chip is introduced in detail, and the correctness of the design is verified through functional simulation. After the process of circuit synthesis and post-simulation, the chip layout design is finally completed. After practice, the chip can effectively improve the data transmission efficiency and system stability, and the research results will help promote the development of the localization of the chip.
{"title":"Design of Communication Controller Chip Based on FlexRay Bus","authors":"Xiaofeng Yang, Yaling Su","doi":"10.1145/3573942.3573977","DOIUrl":"https://doi.org/10.1145/3573942.3573977","url":null,"abstract":"As a new generation of automotive bus, the Flex Ray Alliance includes the largest automotive industry and the most influential. The Flex Ray bus has a very wide range of applications [1], which can promote the development of future automotive electronic systems. The MFR4310 stand-alone FlexRay controller enables easy integration of FlexRay into MCU-based applications and complies with the FlexRay Alliance specification. Driven by the wave of chip localization, in order to solve the problem of insufficient self-sufficiency of China's automotive-grade chips, a communication controller chip based on FlexRay bus was designed, which is fully compatible with MFR4310. In this paper, the implementation of each module of the chip is introduced in detail, and the correctness of the design is verified through functional simulation. After the process of circuit synthesis and post-simulation, the chip layout design is finally completed. After practice, the chip can effectively improve the data transmission efficiency and system stability, and the research results will help promote the development of the localization of the chip.","PeriodicalId":103293,"journal":{"name":"Proceedings of the 2022 5th International Conference on Artificial Intelligence and Pattern Recognition","volume":"21 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115126735","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Shaoshuai Lu, Long Chen, Wenjing Wang, Cai Xu, Wei Zhao, Ziyu Guan, Guangyue Lu
The lack of labeled data is the bottleneck restricting deep text classification algorithm. State-of-the-art for most existing deep text classification methods follow the two-step transfer learning paradigm: pre-training a large model on an auxiliary task, and then fine-tuning the model on a labeled data. Their shortcoming is the high cost of training. To reduce training costs as well as alleviate the need for labeled data, we present a novel simple Semi-Supervised Joint Learning (SSJL) framework for few-shot text classification that captures the rich text semantics from large user-tagged data (referred to as weakly-labeled data) with noisy labels while also learning correct category distributions in small labeled data. We refine the contrastive loss function to better exploit inter-class contrastive patterns, making contrastive learning more applicable to the weakly-labeled setting. Besides, an appropriate temperature hyper-parameter can improve model robustness under label noise. The experimental results on four real-world datasets show that our approach outperforms the other baseline methods. Moreover, SSJL significantly boosts the deep models’ performance with only 0.5% (i.e. 32 samples) of the labeled data, showing its robustness in the data sparsity scenario.
{"title":"A Simple Semi-Supervised Joint Learning Framework for Few-shot Text Classification","authors":"Shaoshuai Lu, Long Chen, Wenjing Wang, Cai Xu, Wei Zhao, Ziyu Guan, Guangyue Lu","doi":"10.1145/3573942.3573945","DOIUrl":"https://doi.org/10.1145/3573942.3573945","url":null,"abstract":"The lack of labeled data is the bottleneck restricting deep text classification algorithm. State-of-the-art for most existing deep text classification methods follow the two-step transfer learning paradigm: pre-training a large model on an auxiliary task, and then fine-tuning the model on a labeled data. Their shortcoming is the high cost of training. To reduce training costs as well as alleviate the need for labeled data, we present a novel simple Semi-Supervised Joint Learning (SSJL) framework for few-shot text classification that captures the rich text semantics from large user-tagged data (referred to as weakly-labeled data) with noisy labels while also learning correct category distributions in small labeled data. We refine the contrastive loss function to better exploit inter-class contrastive patterns, making contrastive learning more applicable to the weakly-labeled setting. Besides, an appropriate temperature hyper-parameter can improve model robustness under label noise. The experimental results on four real-world datasets show that our approach outperforms the other baseline methods. Moreover, SSJL significantly boosts the deep models’ performance with only 0.5% (i.e. 32 samples) of the labeled data, showing its robustness in the data sparsity scenario.","PeriodicalId":103293,"journal":{"name":"Proceedings of the 2022 5th International Conference on Artificial Intelligence and Pattern Recognition","volume":"114 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133929478","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
With the rapid development of the Internet of Things, wireless sensor network, one of its important supporting technologies, has attracted more and more attention. We will work in the low duty cycle wireless sensor network, called low duty cycle wireless sensor network. Neighbor discovery is the most initial but essential work in low duty cycle wireless sensor networks. Although some neighbor discovery algorithms can also achieve neighbor discovery, the average discovery delay is long, and it is difficult to achieve the ability to respond quickly. How to make the nodes in the network quickly realize neighbor discovery is a difficult problem in current research. This paper proposes a group-based fast-response neighbor discovery algorithm (GBFR, in short). At the beginning of the time period, the nodes search for their neighbors by sending a short beacon message, so that the nodes group in pairs. By exchanging neighbor work schedules, nodes know ahead of time some other grouped potential neighbors. Combining the relative distance-based algorithm and node movement, it can selectively recommend suitable neighbors so that nodes can wake up actively and determine whether they are neighbors, thereby speeding up neighbor discovery, reducing communication energy consumption, and improving network life. In this paper, a large number of simulation experiments show that the algorithm has achieved good results in reducing the discovery delay and network energy consumption.
{"title":"A Fast Response Neighbor Discovery Algorithm in Low-Duty-Cycle Mobile Sensor Networks","authors":"Anquan Zhang, Dongming Xu","doi":"10.1145/3573942.3573984","DOIUrl":"https://doi.org/10.1145/3573942.3573984","url":null,"abstract":"With the rapid development of the Internet of Things, wireless sensor network, one of its important supporting technologies, has attracted more and more attention. We will work in the low duty cycle wireless sensor network, called low duty cycle wireless sensor network. Neighbor discovery is the most initial but essential work in low duty cycle wireless sensor networks. Although some neighbor discovery algorithms can also achieve neighbor discovery, the average discovery delay is long, and it is difficult to achieve the ability to respond quickly. How to make the nodes in the network quickly realize neighbor discovery is a difficult problem in current research. This paper proposes a group-based fast-response neighbor discovery algorithm (GBFR, in short). At the beginning of the time period, the nodes search for their neighbors by sending a short beacon message, so that the nodes group in pairs. By exchanging neighbor work schedules, nodes know ahead of time some other grouped potential neighbors. Combining the relative distance-based algorithm and node movement, it can selectively recommend suitable neighbors so that nodes can wake up actively and determine whether they are neighbors, thereby speeding up neighbor discovery, reducing communication energy consumption, and improving network life. In this paper, a large number of simulation experiments show that the algorithm has achieved good results in reducing the discovery delay and network energy consumption.","PeriodicalId":103293,"journal":{"name":"Proceedings of the 2022 5th International Conference on Artificial Intelligence and Pattern Recognition","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131099435","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Named Entity Recognition (NER) is a subtask of natural language processing. Its accuracy is crucial for downstream tasks. In Chinese NER, word information is often added to enhance the semantic and boundary information of Chinese words, but these methods ignore the radical information of Chinese characters. This paper propose a multi-feature fusion model(MFFM) for Chinese NER. First, the input sequences are exported to the BERT layer, the word embedding layer and the radical embedding layer respectively; then the above three layer output are combined together as input of the Bidirectional Long Short-Term Memory(BiLSTM) layer to model the contextual information; finally annotate the sequence with conditional random field. The proposed model not only avoids the import of complex structures, but also effectively captures the character features of the context, thus improves the recognition performance. The experimental results show that the F1 value of MFFM reaches 71.02% on the Weibo dataset, which is 3.12% higher than that of the BERT model, and 82.78% on the OntoNotes4.0 dataset, which is 0.85% higher than that of the BERT model.
{"title":"A Chinese Named Entity Recognition Method Fusing Word and Radical Features","authors":"Shan Deng, Kai-Biao Lin, Ping Lu","doi":"10.1145/3573942.3574055","DOIUrl":"https://doi.org/10.1145/3573942.3574055","url":null,"abstract":"Named Entity Recognition (NER) is a subtask of natural language processing. Its accuracy is crucial for downstream tasks. In Chinese NER, word information is often added to enhance the semantic and boundary information of Chinese words, but these methods ignore the radical information of Chinese characters. This paper propose a multi-feature fusion model(MFFM) for Chinese NER. First, the input sequences are exported to the BERT layer, the word embedding layer and the radical embedding layer respectively; then the above three layer output are combined together as input of the Bidirectional Long Short-Term Memory(BiLSTM) layer to model the contextual information; finally annotate the sequence with conditional random field. The proposed model not only avoids the import of complex structures, but also effectively captures the character features of the context, thus improves the recognition performance. The experimental results show that the F1 value of MFFM reaches 71.02% on the Weibo dataset, which is 3.12% higher than that of the BERT model, and 82.78% on the OntoNotes4.0 dataset, which is 0.85% higher than that of the BERT model.","PeriodicalId":103293,"journal":{"name":"Proceedings of the 2022 5th International Conference on Artificial Intelligence and Pattern Recognition","volume":"60 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131793705","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
As a privacy protection method with strict mathematical definition, differential privacy has been widely used in various fields of data mining including clustering algorithm. However, the traditional differential privacy k-means algorithm is sensitive to the selection of initial value, and the allocation of privacy budget is relatively single, which reduces the availability of the algorithm. In order to further improve the availability of the differential privacy K-means algorithm, this paper proposes a privacy budget allocation method combining error analysis to optimize algorithm iteration times and merge clustering, and carries out theoretical analysis and experimental verification at the same time. The results show that the algorithm not only satisfies the definition of differential privacy, but also improves the availability of clustering effectively.
{"title":"A Differential Privacy K-Means Algorithm for Improving Privacy Budget Allocation","authors":"Sen Liu, Jianhua Liu","doi":"10.1145/3573942.3573957","DOIUrl":"https://doi.org/10.1145/3573942.3573957","url":null,"abstract":"As a privacy protection method with strict mathematical definition, differential privacy has been widely used in various fields of data mining including clustering algorithm. However, the traditional differential privacy k-means algorithm is sensitive to the selection of initial value, and the allocation of privacy budget is relatively single, which reduces the availability of the algorithm. In order to further improve the availability of the differential privacy K-means algorithm, this paper proposes a privacy budget allocation method combining error analysis to optimize algorithm iteration times and merge clustering, and carries out theoretical analysis and experimental verification at the same time. The results show that the algorithm not only satisfies the definition of differential privacy, but also improves the availability of clustering effectively.","PeriodicalId":103293,"journal":{"name":"Proceedings of the 2022 5th International Conference on Artificial Intelligence and Pattern Recognition","volume":"81 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116792254","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This paper presents a prediction model for radar antenna corrosion based on BP neural network optimized by genetic algorithm. The initial connection weights and thresholds of the network model are optimized by the genetic algorithm, then the BP network optimized by the genetic algorithm is designed, and the method is validated by simulation using the prediction of radar whole machine corrosion as an example. The experimental results show that the prediction of radar antenna corrosion based on GA-BP meets the accuracy requirements of radar antenna corrosion prediction.
{"title":"Research on Radar Corrosion Prediction Model Based on BP Neural Network Optimized by Genetic Algorithm","authors":"Duanquan Fan, Lei Yin, Longwen Shen","doi":"10.1145/3573942.3573955","DOIUrl":"https://doi.org/10.1145/3573942.3573955","url":null,"abstract":"This paper presents a prediction model for radar antenna corrosion based on BP neural network optimized by genetic algorithm. The initial connection weights and thresholds of the network model are optimized by the genetic algorithm, then the BP network optimized by the genetic algorithm is designed, and the method is validated by simulation using the prediction of radar whole machine corrosion as an example. The experimental results show that the prediction of radar antenna corrosion based on GA-BP meets the accuracy requirements of radar antenna corrosion prediction.","PeriodicalId":103293,"journal":{"name":"Proceedings of the 2022 5th International Conference on Artificial Intelligence and Pattern Recognition","volume":"34 1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125100820","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","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：