Pub Date : 2022-07-15DOI: 10.1109/icaci55529.2022.9837584
Ming Yan, Pingyue Zhang, Yinlin Hao, Mingxin Kang, Yuhu Wu
In this paper, the parameters tunning problem for a nonlinear active disturbance rejection controller (NLADRC), which is applied to control a special air path system, including the exhaust gas recirculation and variable geometry turbine, in diesel engine. One of the main challenges in design NLADRC for a practical system is to tunning the parameters NLADRC, which are too many and affect each others. In this paper, based on continuous action reinforcement learning automata (CARLA), a NLADRC-CARLA parameter tunning algorithm is proposed. This algorithm can automatically learn the parameters satisfying the control performance. To verity the effectiveness of the proposed algorithm, simulation results are given for a variable geometry turbine and exhaust gas recirculation (VGT-EGR) system in a diesel engine.
{"title":"Parameter Tunning of NLADRC for VGT-EGR System in Diesel Engine Based on CARLA Algorithm","authors":"Ming Yan, Pingyue Zhang, Yinlin Hao, Mingxin Kang, Yuhu Wu","doi":"10.1109/icaci55529.2022.9837584","DOIUrl":"https://doi.org/10.1109/icaci55529.2022.9837584","url":null,"abstract":"In this paper, the parameters tunning problem for a nonlinear active disturbance rejection controller (NLADRC), which is applied to control a special air path system, including the exhaust gas recirculation and variable geometry turbine, in diesel engine. One of the main challenges in design NLADRC for a practical system is to tunning the parameters NLADRC, which are too many and affect each others. In this paper, based on continuous action reinforcement learning automata (CARLA), a NLADRC-CARLA parameter tunning algorithm is proposed. This algorithm can automatically learn the parameters satisfying the control performance. To verity the effectiveness of the proposed algorithm, simulation results are given for a variable geometry turbine and exhaust gas recirculation (VGT-EGR) system in a diesel engine.","PeriodicalId":412347,"journal":{"name":"2022 14th International Conference on Advanced Computational Intelligence (ICACI)","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131995475","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}
Pub Date : 2022-07-15DOI: 10.1109/icaci55529.2022.9837710
Cong Teng, Liyan Song
Just-In-Time Software Defect Predict (JIT-SDP) has been a popular research topic in the literature of software engineering. In many practical scenarios, software engineers would prefer to pursue the best detection of defect-inducing software changes under the concern of a given false alarm tolerance. However, there have been only two related studies in the Machine Learning (ML) community that are capable of tackling this constraint optimization problem. This paper aims to study how can we utilize the existing ML methods for addressing the research problem in JIT-SDP and how well do they perform on it. Considering the fact that the objective and the constraint are not differentiable, a Differential Evolution (DE) algorithm is by nature suitable for tackling this research problem. Thus, this paper also aims to investigate how can we propose a novel DE algorithm to better address the constraint optimization problem in JIT-SDP. With these aims in mind, this paper adapts the ML methods with a spared validation set to facilitate the constraint learning process, and it also proposes an advanced DE algorithm with an adaptive constraint to pursue the best detection of the positive class under a given false alarm. Experimental results with 10 real-world data sets from the domain of software defect prediction demonstrate that our proposed DE based approach can achieve generally better performance on the constraint optimization problem, deriving better classification models in terms of both objective and the constraint.
{"title":"In Pursuit of the Best Detection of Positive Data Under User’s Concern on False Alarm","authors":"Cong Teng, Liyan Song","doi":"10.1109/icaci55529.2022.9837710","DOIUrl":"https://doi.org/10.1109/icaci55529.2022.9837710","url":null,"abstract":"Just-In-Time Software Defect Predict (JIT-SDP) has been a popular research topic in the literature of software engineering. In many practical scenarios, software engineers would prefer to pursue the best detection of defect-inducing software changes under the concern of a given false alarm tolerance. However, there have been only two related studies in the Machine Learning (ML) community that are capable of tackling this constraint optimization problem. This paper aims to study how can we utilize the existing ML methods for addressing the research problem in JIT-SDP and how well do they perform on it. Considering the fact that the objective and the constraint are not differentiable, a Differential Evolution (DE) algorithm is by nature suitable for tackling this research problem. Thus, this paper also aims to investigate how can we propose a novel DE algorithm to better address the constraint optimization problem in JIT-SDP. With these aims in mind, this paper adapts the ML methods with a spared validation set to facilitate the constraint learning process, and it also proposes an advanced DE algorithm with an adaptive constraint to pursue the best detection of the positive class under a given false alarm. Experimental results with 10 real-world data sets from the domain of software defect prediction demonstrate that our proposed DE based approach can achieve generally better performance on the constraint optimization problem, deriving better classification models in terms of both objective and the constraint.","PeriodicalId":412347,"journal":{"name":"2022 14th International Conference on Advanced Computational Intelligence (ICACI)","volume":"26 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123540850","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}
Pub Date : 2022-07-15DOI: 10.1109/icaci55529.2022.9837747
Jiejie Chen, Boshan Chen, Z. Zeng
In this paper, firstly, a novel dynamic event-triggered adjustable control protocol is proposed, which can be distributed, centralized and mixed by tuning one parameter, where the dynamic event-triggered mechanism includes many existing (dynamic and static) event-triggering mechanisms as special cases. Then with this control protocol, we deal with the leader-follower consensus problem for multi-agent systems. It is shown that the multi-agent system do not exhibit Zeno behavior, and can achieve leader-follower consensus as well as under this control protocol. Finally, an algorithm is provided to avoid continuous communication when the dynamic event-triggering mechanism is implemented. In addition, a numerical example is given to illustrate the validity of the obtained results and the advantage of the proposed control protocol.
{"title":"Leader-Following Consensus of Linear Multi-Agent Systems via Dynamic Event-Triggered Adjustable Control Protocol","authors":"Jiejie Chen, Boshan Chen, Z. Zeng","doi":"10.1109/icaci55529.2022.9837747","DOIUrl":"https://doi.org/10.1109/icaci55529.2022.9837747","url":null,"abstract":"In this paper, firstly, a novel dynamic event-triggered adjustable control protocol is proposed, which can be distributed, centralized and mixed by tuning one parameter, where the dynamic event-triggered mechanism includes many existing (dynamic and static) event-triggering mechanisms as special cases. Then with this control protocol, we deal with the leader-follower consensus problem for multi-agent systems. It is shown that the multi-agent system do not exhibit Zeno behavior, and can achieve leader-follower consensus as well as under this control protocol. Finally, an algorithm is provided to avoid continuous communication when the dynamic event-triggering mechanism is implemented. In addition, a numerical example is given to illustrate the validity of the obtained results and the advantage of the proposed control protocol.","PeriodicalId":412347,"journal":{"name":"2022 14th International Conference on Advanced Computational Intelligence (ICACI)","volume":"46 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121588081","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}
Pub Date : 2022-07-15DOI: 10.1109/icaci55529.2022.9837682
Yimei Zhang
In order to solve the problems of slow convergence speed and easy to fall into local optimum in solving the robot path planning problem, this paper improves the basic genetic algorithm. This paper introduces the artificial potential field method to initialize the population, and proposes an adaptive selection method based on the evaluation of the degree of population diversity. The adaptive crossover probability and mutation probability are designed to improve the algorithm solution quality, and multiple simulations are carried out in the grid environment to further prove the feasibility and effectiveness of the algorithm.
{"title":"Research on Robot Path Planning Based on Improved Genetic Algorithm","authors":"Yimei Zhang","doi":"10.1109/icaci55529.2022.9837682","DOIUrl":"https://doi.org/10.1109/icaci55529.2022.9837682","url":null,"abstract":"In order to solve the problems of slow convergence speed and easy to fall into local optimum in solving the robot path planning problem, this paper improves the basic genetic algorithm. This paper introduces the artificial potential field method to initialize the population, and proposes an adaptive selection method based on the evaluation of the degree of population diversity. The adaptive crossover probability and mutation probability are designed to improve the algorithm solution quality, and multiple simulations are carried out in the grid environment to further prove the feasibility and effectiveness of the algorithm.","PeriodicalId":412347,"journal":{"name":"2022 14th International Conference on Advanced Computational Intelligence (ICACI)","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127977718","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}
Sentiment analysis has become one of the most active topics in education research. So far, however, there has been little discussion about the recent application of sentiment analysis for Chinese MOOC reviews. Therefore, this paper sheds light on some fine-grained sentiment analysis technology to benefit the current students and education practitioners. Firstly, we focus on extracting aspect terms associated with the course via dependency parsing and sentiment word lexicons. Secondly, we categorize the aspect terms with the Naive Bayes. Experimental results effectively demonstrate that the proposed approach and refine the granularity of sentiment categories in higher education. This paper makes sentiment analysis possible to increase students’ learning retention and improve teachers’ performance in online teaching.
{"title":"Aspect Term Extraction and Categorization for Chinese MOOC Reviews","authors":"Kangan Zhou, Guangmin Li, Jiejie Chen, Wenjing Chen, Xinhua Xu, Xiaowei Yan","doi":"10.1109/icaci55529.2022.9837511","DOIUrl":"https://doi.org/10.1109/icaci55529.2022.9837511","url":null,"abstract":"Sentiment analysis has become one of the most active topics in education research. So far, however, there has been little discussion about the recent application of sentiment analysis for Chinese MOOC reviews. Therefore, this paper sheds light on some fine-grained sentiment analysis technology to benefit the current students and education practitioners. Firstly, we focus on extracting aspect terms associated with the course via dependency parsing and sentiment word lexicons. Secondly, we categorize the aspect terms with the Naive Bayes. Experimental results effectively demonstrate that the proposed approach and refine the granularity of sentiment categories in higher education. This paper makes sentiment analysis possible to increase students’ learning retention and improve teachers’ performance in online teaching.","PeriodicalId":412347,"journal":{"name":"2022 14th International Conference on Advanced Computational Intelligence (ICACI)","volume":"29 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114237957","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}
Pub Date : 2022-07-15DOI: 10.1109/icaci55529.2022.9837610
Chenxi Song, Sitian Qin, Jiqiang Feng
This paper explores the multistability of S-asymptotically $omega$-periodic solutions for fractional-order neural networks with time variable delays (FVDNNs). Benefited from the geometrical configuration of the nonlinear and non-monotonic activation function, we prove the coexistence of $(K+1)^{n}$ S-asymptotically $omega$-periodic solutions with multiple asymptotical stability, where K is a positive integer. In contrast to the previous works, the obtained results extensively raise the amount of S-asymptotically $omega$-periodic solutions of FVDNNs in this paper. Besides, two numerical examples are shown to illustrate the feasibility of obtained results.
本文研究了具有时变时滞的分数阶神经网络(FVDNNs)的s -渐近周期解的多重稳定性。利用非线性非单调激活函数的几何构型，证明了具有多重渐近稳定性的$(K+1)^{n}$ s -周期解的共存性，其中K为正整数。与以往的工作相比，本文得到的结果广泛地提高了fvdnn的s -渐近$ ω $周期解的数量。并通过两个数值算例说明了所得结果的可行性。
{"title":"Multistability of S-Asymptotically ω-Periodic Solutions for Fractional-Order Neural Networks with Time Variable Delays","authors":"Chenxi Song, Sitian Qin, Jiqiang Feng","doi":"10.1109/icaci55529.2022.9837610","DOIUrl":"https://doi.org/10.1109/icaci55529.2022.9837610","url":null,"abstract":"This paper explores the multistability of S-asymptotically $omega$-periodic solutions for fractional-order neural networks with time variable delays (FVDNNs). Benefited from the geometrical configuration of the nonlinear and non-monotonic activation function, we prove the coexistence of $(K+1)^{n}$ S-asymptotically $omega$-periodic solutions with multiple asymptotical stability, where K is a positive integer. In contrast to the previous works, the obtained results extensively raise the amount of S-asymptotically $omega$-periodic solutions of FVDNNs in this paper. Besides, two numerical examples are shown to illustrate the feasibility of obtained results.","PeriodicalId":412347,"journal":{"name":"2022 14th International Conference on Advanced Computational Intelligence (ICACI)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130759968","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}
Pub Date : 2022-07-15DOI: 10.1109/icaci55529.2022.9837732
Rui Jiao, Sai Li, Zhixia Ding, Guan Wang
In this paper, a rolling bearing fault diagnosis method based on smoothness priors approach considering bearing fault location and damage degree is proposed. Firstly, smoothness priors approach is used to adaptively decompose the bearing vibration signals to obtain the trend and detrended terms; then the combined permutation entropy and energy entropy are used to extract the fault features from the trend and detrended terms to obtain the information entropy feature vectors; finally, the information entropy feature vectors are input to the support vector classifier of sine cosine algorithm. This method is applied to the experimental data of rolling bearing. The analysis results show that the diagnosis effect of using the combination of permutation entropy and energy entropy to extract fault features is better than using only permutation entropy to extract fault features when the bearing fault location and damage degree are considered at the same time.
{"title":"Rolling Bearing Fault Diagnosis Considering Fault Location and Damage Degree Based on Smoothness Priors Approach","authors":"Rui Jiao, Sai Li, Zhixia Ding, Guan Wang","doi":"10.1109/icaci55529.2022.9837732","DOIUrl":"https://doi.org/10.1109/icaci55529.2022.9837732","url":null,"abstract":"In this paper, a rolling bearing fault diagnosis method based on smoothness priors approach considering bearing fault location and damage degree is proposed. Firstly, smoothness priors approach is used to adaptively decompose the bearing vibration signals to obtain the trend and detrended terms; then the combined permutation entropy and energy entropy are used to extract the fault features from the trend and detrended terms to obtain the information entropy feature vectors; finally, the information entropy feature vectors are input to the support vector classifier of sine cosine algorithm. This method is applied to the experimental data of rolling bearing. The analysis results show that the diagnosis effect of using the combination of permutation entropy and energy entropy to extract fault features is better than using only permutation entropy to extract fault features when the bearing fault location and damage degree are considered at the same time.","PeriodicalId":412347,"journal":{"name":"2022 14th International Conference on Advanced Computational Intelligence (ICACI)","volume":"519 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123117052","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}
Pub Date : 2022-07-15DOI: 10.1109/icaci55529.2022.9837549
Xiaoping Zhang, Tianhang Yang, Li Wang, Zhonghe He, Shida Liu
Deep learning is an important branch of neural networks, which has high accuracy in classification and regression problems, and has been widely used. However, its performance is greatly affected by the parameters. In this paper, an improved particle swarm algorithm named as PSO-C is proposed to automatically train the parameters of the feedforward neural networks. In the proposed algorithm, the curiosity factor is introduced to divide the particles into two categories with different curiosity characteristics so as to improve the exploration ability and information mining ability of the particle swarms. At the same time, a chaotic factor is also introduced to avoid the local optimum problem during the neural network’s training. The simulation results show that the PSO-C has better optimization effect on the whole.
{"title":"An Improved Particle Swarm Optimization Algorithm for Parameters Optimizing of Feedforward Neural Networks","authors":"Xiaoping Zhang, Tianhang Yang, Li Wang, Zhonghe He, Shida Liu","doi":"10.1109/icaci55529.2022.9837549","DOIUrl":"https://doi.org/10.1109/icaci55529.2022.9837549","url":null,"abstract":"Deep learning is an important branch of neural networks, which has high accuracy in classification and regression problems, and has been widely used. However, its performance is greatly affected by the parameters. In this paper, an improved particle swarm algorithm named as PSO-C is proposed to automatically train the parameters of the feedforward neural networks. In the proposed algorithm, the curiosity factor is introduced to divide the particles into two categories with different curiosity characteristics so as to improve the exploration ability and information mining ability of the particle swarms. At the same time, a chaotic factor is also introduced to avoid the local optimum problem during the neural network’s training. The simulation results show that the PSO-C has better optimization effect on the whole.","PeriodicalId":412347,"journal":{"name":"2022 14th International Conference on Advanced Computational Intelligence (ICACI)","volume":"31 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123945137","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}
Pub Date : 2022-07-15DOI: 10.1109/icaci55529.2022.9837556
Hao Su, Shiwu Yang, Chang Liu, Haiwei Liu
The unstructured text of railway signal equipment failure records important information such as the failure cause and failure phenomenon of the signal equipment. Most of them are stored in Word, Excel, etc. The traditional technology cannot explore the important value contained in the text data. In order to convert the analysis of the fault causes of the signal equipment recorded in the text into knowledge that can serve fault diagnosis, this paper uses the BiLSTM+CRF model to realize named entity recognition and analyzes 638 fault texts of railway signal equipment in a railway field from 2021 to 2022. The accuracy of the model reaches 83.38%, which shows that the named entity recognition model of railway signal fault equipment has a high evaluation standard and can be applied to the extraction of signal equipment fault entities based on text mining.
{"title":"Research on Named Entity Recognition in Fault Text of Railway Signal Equipment","authors":"Hao Su, Shiwu Yang, Chang Liu, Haiwei Liu","doi":"10.1109/icaci55529.2022.9837556","DOIUrl":"https://doi.org/10.1109/icaci55529.2022.9837556","url":null,"abstract":"The unstructured text of railway signal equipment failure records important information such as the failure cause and failure phenomenon of the signal equipment. Most of them are stored in Word, Excel, etc. The traditional technology cannot explore the important value contained in the text data. In order to convert the analysis of the fault causes of the signal equipment recorded in the text into knowledge that can serve fault diagnosis, this paper uses the BiLSTM+CRF model to realize named entity recognition and analyzes 638 fault texts of railway signal equipment in a railway field from 2021 to 2022. The accuracy of the model reaches 83.38%, which shows that the named entity recognition model of railway signal fault equipment has a high evaluation standard and can be applied to the extraction of signal equipment fault entities based on text mining.","PeriodicalId":412347,"journal":{"name":"2022 14th International Conference on Advanced Computational Intelligence (ICACI)","volume":"10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124478258","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}
Cage aquaculture is one of the important types of marine aquaculture, and reasonable monitoring can achieve sustainable and stable development. Using the Synthetic Aperture Radar (SAR) to realize the extraction of cage aquaculture is significant. The convolutional neural networks (CNN) extract cage aquaculture by learning semantic information from deep features. However, training CNN usually needs a large number of labeled samples. Unsupervised learning is difficult to discover the semantic information of aquaculture due to the speckle noise in SAR images. In this article, an invariant information differentiable feature clustering network (IIDFCN) is proposed to enhance spatial continuity and reduce the influence of speckle noise. The pseudo-labels are obtained by a differentiable function processing the deep features of network output. The network parameters are updated by back-propagation, and the deep features and pseudo-labels are alternately and jointly optimized. In addition, in order to obtain reasonable spatial continuity constraints, an invariant information loss is introduced into the global loss function. The IIDFCN solves the problem of needing a large number of labels in the extraction of SAR aquaculture and implements the unsupervised deep network learning of cage aquaculture semantic information. The experiments test the method on a cage aquaculture data set from the Sanduao area, which shows the approach to be effective.
{"title":"Unsupervised Segmentation of Cage Aquaculture in SAR Images Based on Invariant Information","authors":"Jianlin Zhou, Chu Chu, Gongwen Zhou, Xinzhe Wang, Kelin Wang, Jianchao Fan","doi":"10.1109/icaci55529.2022.9837739","DOIUrl":"https://doi.org/10.1109/icaci55529.2022.9837739","url":null,"abstract":"Cage aquaculture is one of the important types of marine aquaculture, and reasonable monitoring can achieve sustainable and stable development. Using the Synthetic Aperture Radar (SAR) to realize the extraction of cage aquaculture is significant. The convolutional neural networks (CNN) extract cage aquaculture by learning semantic information from deep features. However, training CNN usually needs a large number of labeled samples. Unsupervised learning is difficult to discover the semantic information of aquaculture due to the speckle noise in SAR images. In this article, an invariant information differentiable feature clustering network (IIDFCN) is proposed to enhance spatial continuity and reduce the influence of speckle noise. The pseudo-labels are obtained by a differentiable function processing the deep features of network output. The network parameters are updated by back-propagation, and the deep features and pseudo-labels are alternately and jointly optimized. In addition, in order to obtain reasonable spatial continuity constraints, an invariant information loss is introduced into the global loss function. The IIDFCN solves the problem of needing a large number of labels in the extraction of SAR aquaculture and implements the unsupervised deep network learning of cage aquaculture semantic information. The experiments test the method on a cage aquaculture data set from the Sanduao area, which shows the approach to be effective.","PeriodicalId":412347,"journal":{"name":"2022 14th International Conference on Advanced Computational Intelligence (ICACI)","volume":"55 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124836559","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：