Pub Date : 2022-07-15DOI: 10.1109/icaci55529.2022.9837756
Shuyang Wang, Huiying Liu, Yuhuai Chen, Zengyang Li
The Saihanba ecological model has achieved brilliant success in environmental protection. Given that desertification and environmental pollution are serious problems in many areas, it is valuable to copy the success of the Saihanba Forest to other areas in need of China. In this paper, we first proposes an ecological protected area (EPA) evaluation method focusing on (1) the impact of the EPA on the environment before and after the establishment of the EPA, and (2) the environmental influence of the EPA on surrounding areas. Specifically, the principal component analysis method is used to evaluate the environmental impact of Saihanba before and after its restoration. Taking the impact of Saihanba on sandstorms in North China as an example, the multiple linear regression method is used to establish the impact model in the surrounding areas. According to the impact model, the construction of Saihanba reduces about 9.8 days of sand-dust weather in North China every year. In addition, this paper also proposes an EPA site selection method. To be specific, this paper explores how to promote the Saihanba ecological model to establish EPA in China. Through cluster analysis, Ordos and Alxa League in Inner Mongolia as well as Yulin in Shaanxi are selected as the sites to establish ecological reserves. To balance ecological, economic, and industrial land and to ensure that target areas have enough land, the area (size) of the EPA is estimated according to the Markov model. Besides, the impact on China’s carbon neutrality goals is assessed and the contribution rates of Ordos, Alxa, and Yulin to the reduction of national carbon emissions are 2.8%0,1.1%0, and 1.7%0,, respectively.
{"title":"Ecological Protected Area Evaluation and Site Selection Methods Based on the Saihanba Ecological Model","authors":"Shuyang Wang, Huiying Liu, Yuhuai Chen, Zengyang Li","doi":"10.1109/icaci55529.2022.9837756","DOIUrl":"https://doi.org/10.1109/icaci55529.2022.9837756","url":null,"abstract":"The Saihanba ecological model has achieved brilliant success in environmental protection. Given that desertification and environmental pollution are serious problems in many areas, it is valuable to copy the success of the Saihanba Forest to other areas in need of China. In this paper, we first proposes an ecological protected area (EPA) evaluation method focusing on (1) the impact of the EPA on the environment before and after the establishment of the EPA, and (2) the environmental influence of the EPA on surrounding areas. Specifically, the principal component analysis method is used to evaluate the environmental impact of Saihanba before and after its restoration. Taking the impact of Saihanba on sandstorms in North China as an example, the multiple linear regression method is used to establish the impact model in the surrounding areas. According to the impact model, the construction of Saihanba reduces about 9.8 days of sand-dust weather in North China every year. In addition, this paper also proposes an EPA site selection method. To be specific, this paper explores how to promote the Saihanba ecological model to establish EPA in China. Through cluster analysis, Ordos and Alxa League in Inner Mongolia as well as Yulin in Shaanxi are selected as the sites to establish ecological reserves. To balance ecological, economic, and industrial land and to ensure that target areas have enough land, the area (size) of the EPA is estimated according to the Markov model. Besides, the impact on China’s carbon neutrality goals is assessed and the contribution rates of Ordos, Alxa, and Yulin to the reduction of national carbon emissions are 2.8%0,1.1%0, and 1.7%0,, respectively.","PeriodicalId":412347,"journal":{"name":"2022 14th International Conference on Advanced Computational Intelligence (ICACI)","volume":"81 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":"126277163","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/icaci55529.2022.9837527
Guanghu Kuang, Jianchao Fan, Jun Wang
Floating raft aquaculture is an effective method in the China coastal sea. Compared to synthetic aperture radar (SAR), polarimetric synthetic aperture radar (PolSAR) can obtain more echo information and enhance the ability of imaging radar to get target information. MDOAU-net has been successful in SAR images’ marine aquaculture detection encouraging researchers to explore the performance of PolSAR data in MDOAU-net. However, MDOAU-net did not consider that PolSAR data have more multi-scattering information of objects, and SAR data only have intensity information of objects. Moreover, compared to SAR data, PolSAR data has fewer speckle noises. So, this paper proposes a new model called PMVOAU-net, which is faster and more effective for PolSAR image segmentation than MDOAU-net. Adopting the Freeman decomposition, getting pseudo-color images of scattering characteristics fusion and three components scattering images. Experiments on PolSAR images substantiate the effectiveness of the proposed approach.
{"title":"PolSAR Marine Aquaculture Detection Based on Fast PMVOAU-net","authors":"Guanghu Kuang, Jianchao Fan, Jun Wang","doi":"10.1109/icaci55529.2022.9837527","DOIUrl":"https://doi.org/10.1109/icaci55529.2022.9837527","url":null,"abstract":"Floating raft aquaculture is an effective method in the China coastal sea. Compared to synthetic aperture radar (SAR), polarimetric synthetic aperture radar (PolSAR) can obtain more echo information and enhance the ability of imaging radar to get target information. MDOAU-net has been successful in SAR images’ marine aquaculture detection encouraging researchers to explore the performance of PolSAR data in MDOAU-net. However, MDOAU-net did not consider that PolSAR data have more multi-scattering information of objects, and SAR data only have intensity information of objects. Moreover, compared to SAR data, PolSAR data has fewer speckle noises. So, this paper proposes a new model called PMVOAU-net, which is faster and more effective for PolSAR image segmentation than MDOAU-net. Adopting the Freeman decomposition, getting pseudo-color images of scattering characteristics fusion and three components scattering images. Experiments on PolSAR images substantiate the effectiveness of the proposed approach.","PeriodicalId":412347,"journal":{"name":"2022 14th International Conference on Advanced Computational Intelligence (ICACI)","volume":"65 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":"133267092","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/icaci55529.2022.9837611
Chunsheng Zhao, Xiukun Wei, Jing Li
In recent years, computer vision-based moving target state detection and motion parameter estimation have become a hot topic. Aiming at the problem that the accurate speed of moving target cannot be calculated from the target tracking results, in this paper, a speed estimation method using Siamese convolutional network and Kalman filtering is proposed and it is validated by single pendulum experimental data. Firstly, an improved Siamese convolutional network is integrated to detect the position of a moving ball in the single pendulum video. Then, these position coordinates are integrated to estimate the speed of the moving ball by using the information from the previous and current frame through a Kalman filter. The experimental results show that this method can achieve a sound on the speed estimation of moving objects in videos.
{"title":"Speed Estimation of Video Target Based on Siamese Convolutional Network and Kalman Filtering","authors":"Chunsheng Zhao, Xiukun Wei, Jing Li","doi":"10.1109/icaci55529.2022.9837611","DOIUrl":"https://doi.org/10.1109/icaci55529.2022.9837611","url":null,"abstract":"In recent years, computer vision-based moving target state detection and motion parameter estimation have become a hot topic. Aiming at the problem that the accurate speed of moving target cannot be calculated from the target tracking results, in this paper, a speed estimation method using Siamese convolutional network and Kalman filtering is proposed and it is validated by single pendulum experimental data. Firstly, an improved Siamese convolutional network is integrated to detect the position of a moving ball in the single pendulum video. Then, these position coordinates are integrated to estimate the speed of the moving ball by using the information from the previous and current frame through a Kalman filter. The experimental results show that this method can achieve a sound on the speed estimation of moving objects in videos.","PeriodicalId":412347,"journal":{"name":"2022 14th International Conference on Advanced Computational Intelligence (ICACI)","volume":"35 4","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114085896","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/icaci55529.2022.9837717
Xiao Nan, Jiahui Yu, Shuai Wang
This paper intends to solve the robust consensus problem of cyber-phycial networks with signed graph under cyber-attacks. The conmunication topology is considered to be directed. First, A fully distributed robust control protocol is presented without uasage of any glabol information, when actuator suffered false dada input attacks. Next, the stalibity is proved by Lyapounov Theory. Finally, a simulation to verify the main result in this paper.
{"title":"Bipartite Consensus of Cyber-phycial Networks against False Data Injection Attacks","authors":"Xiao Nan, Jiahui Yu, Shuai Wang","doi":"10.1109/icaci55529.2022.9837717","DOIUrl":"https://doi.org/10.1109/icaci55529.2022.9837717","url":null,"abstract":"This paper intends to solve the robust consensus problem of cyber-phycial networks with signed graph under cyber-attacks. The conmunication topology is considered to be directed. First, A fully distributed robust control protocol is presented without uasage of any glabol information, when actuator suffered false dada input attacks. Next, the stalibity is proved by Lyapounov Theory. Finally, a simulation to verify the main result in this paper.","PeriodicalId":412347,"journal":{"name":"2022 14th International Conference on Advanced Computational Intelligence (ICACI)","volume":"93 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":"124173190","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/icaci55529.2022.9837706
Zhonghua Chen, Hongkai Wang, F. Cong, Lauri Kettunen
The construction of statistical shape models (SSMs) is an important method in the field of medical image segmentation. Most SSMs are constructed by using traditional modeling methods based on principal component analysis (PCA), which cannot fully present the true deformation ability of models. To solve the insufficient deformation ability of SSMs, we propose a stacked autoencoder (SAE) neural network to construct a multi-resolution multi-organ shape model based on mouse micro-CT images, which can express more linear and non-linear deformations than SSMs based on PCA. The main advantage of this method is that the SAE neural network is simple and flexible and it can learn more deformation modes from training data. We have quantitatively compared the modeling performance of this method with the constructed SSMs based on PCA in terms of model generalization and specificity.
{"title":"Construction of Multi-resolution Multi-organ Shape Model Based on Stacked Autoencoder Neural Network","authors":"Zhonghua Chen, Hongkai Wang, F. Cong, Lauri Kettunen","doi":"10.1109/icaci55529.2022.9837706","DOIUrl":"https://doi.org/10.1109/icaci55529.2022.9837706","url":null,"abstract":"The construction of statistical shape models (SSMs) is an important method in the field of medical image segmentation. Most SSMs are constructed by using traditional modeling methods based on principal component analysis (PCA), which cannot fully present the true deformation ability of models. To solve the insufficient deformation ability of SSMs, we propose a stacked autoencoder (SAE) neural network to construct a multi-resolution multi-organ shape model based on mouse micro-CT images, which can express more linear and non-linear deformations than SSMs based on PCA. The main advantage of this method is that the SAE neural network is simple and flexible and it can learn more deformation modes from training data. We have quantitatively compared the modeling performance of this method with the constructed SSMs based on PCA in terms of model generalization and specificity.","PeriodicalId":412347,"journal":{"name":"2022 14th International Conference on Advanced Computational Intelligence (ICACI)","volume":"2 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":"128066986","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/icaci55529.2022.9837529
Lei Feng, Zongwu Ke, Na Wu
Multimodel deep learning system has attracted more and more attention. The traditional deep learning system mostly focuses on single modal processing and application. However, many applications require various modes to complete a certain task. For example, in the teaching scene, teaching materials are mostly displayed in text mode, video and PPT modes are also used to transfer the content of knowledge points. However, there is often a lack of connection between one mode and another mode, resulting in the fragmentation, complexity and redundancy of knowledge. Based on this consideration, the paper puts forward the design idea and frame of multimodal curriculum knowledge graph based on paddleOCR and DeepKE. Use DeepKE to extract the triple relationship between subject knowledge points and store it in the neo4j graph database, so as to build the knowledge graph of subject knowledge points, then use paddeOCR to identify the text content in the teaching video, generate the video frame description text, use NLP processing technologies such as text similarity to realize the understanding of video segments, and finally link the fine-grained video segments to the text knowledge graph, so as to build the multi-modal curriculum knowledge graph, In order to realize the purpose of intelligent search and intelligent construction of learning link.
{"title":"ModelsKG:A Design and Research on Knowledge Graph of Multimodal Curriculum Based on PaddleOCR and DeepKE","authors":"Lei Feng, Zongwu Ke, Na Wu","doi":"10.1109/icaci55529.2022.9837529","DOIUrl":"https://doi.org/10.1109/icaci55529.2022.9837529","url":null,"abstract":"Multimodel deep learning system has attracted more and more attention. The traditional deep learning system mostly focuses on single modal processing and application. However, many applications require various modes to complete a certain task. For example, in the teaching scene, teaching materials are mostly displayed in text mode, video and PPT modes are also used to transfer the content of knowledge points. However, there is often a lack of connection between one mode and another mode, resulting in the fragmentation, complexity and redundancy of knowledge. Based on this consideration, the paper puts forward the design idea and frame of multimodal curriculum knowledge graph based on paddleOCR and DeepKE. Use DeepKE to extract the triple relationship between subject knowledge points and store it in the neo4j graph database, so as to build the knowledge graph of subject knowledge points, then use paddeOCR to identify the text content in the teaching video, generate the video frame description text, use NLP processing technologies such as text similarity to realize the understanding of video segments, and finally link the fine-grained video segments to the text knowledge graph, so as to build the multi-modal curriculum knowledge graph, In order to realize the purpose of intelligent search and intelligent construction of learning link.","PeriodicalId":412347,"journal":{"name":"2022 14th International Conference on Advanced Computational Intelligence (ICACI)","volume":"330 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":"122744523","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}
Landslide displacement prediction is essential to establishing the early warning system (EWS). To better grasp the landslide evolution process, this paper proposes a novel architecture of variational mode decomposition-Generative Adversarial Network (VMD-GAN) for forecasting the landslide displacement. Firstly, VMD was used to decompose the time series into multiple intrinsic mode functions (IMFs) to extract the internal hidden information of the original series and remove the interference of noise to improve the prediction accuracy of the model. Then, GAN predicts each IMFs. Finally, the predicted results for each IMFs component are added to get the final prediction result. The Baishuihe in the Three Gorges Reservoir was made as an example and displacement data from August 2003 to December 2011 were selected for analysis. Compared with empirical mode decomposition-Generative Adversarial Network(EMDGAN), long short-term memory (LSTM), and temporal convolutional networks (TCN) models, the result has shown that the root means square errors (RMSE) of VMD-GAN in landslide prediction was 3.33 and the correlation coefficient R-square was 0.99, which demonstrated the best prediction accuracy and fitting ability.
{"title":"Combined with Decomposition Algorithm and Generative Adversarial Networks on Landslide Displacement Prediction","authors":"Mengfei Xu, Jiejie Chen, Honggang Yang, Tongfei Xiao","doi":"10.1109/icaci55529.2022.9837779","DOIUrl":"https://doi.org/10.1109/icaci55529.2022.9837779","url":null,"abstract":"Landslide displacement prediction is essential to establishing the early warning system (EWS). To better grasp the landslide evolution process, this paper proposes a novel architecture of variational mode decomposition-Generative Adversarial Network (VMD-GAN) for forecasting the landslide displacement. Firstly, VMD was used to decompose the time series into multiple intrinsic mode functions (IMFs) to extract the internal hidden information of the original series and remove the interference of noise to improve the prediction accuracy of the model. Then, GAN predicts each IMFs. Finally, the predicted results for each IMFs component are added to get the final prediction result. The Baishuihe in the Three Gorges Reservoir was made as an example and displacement data from August 2003 to December 2011 were selected for analysis. Compared with empirical mode decomposition-Generative Adversarial Network(EMDGAN), long short-term memory (LSTM), and temporal convolutional networks (TCN) models, the result has shown that the root means square errors (RMSE) of VMD-GAN in landslide prediction was 3.33 and the correlation coefficient R-square was 0.99, which demonstrated the best prediction accuracy and fitting ability.","PeriodicalId":412347,"journal":{"name":"2022 14th International Conference on Advanced Computational Intelligence (ICACI)","volume":"47 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":"121585506","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/icaci55529.2022.9837693
Yilin Yan, Junwei Sun, Yongxing Ma, J. Yang, Peng Liu, Yanfeng Wang
In this paper, a locally active hyperbolic memristor is proposed which characteristics are analyzed by numerical analysis. The HR-FN neuron model which bidirectional coupled by locally active hyperbolic memristor is constructed. Theoretical analysis and numerical simulation show that the bidirectional coupled neuron model has multiple stability under the influence of locally active memristor. Finally, the equivalent circuit of bidirectional coupled neural network is designed, and its the numerical analysis is verified by Multisim circuit simulation.
{"title":"Dynamical Analysis of HR-FN Neuron Model Bidirectional Coupled by Locally Active Hyperbolic Memristor and Circuit Implementation","authors":"Yilin Yan, Junwei Sun, Yongxing Ma, J. Yang, Peng Liu, Yanfeng Wang","doi":"10.1109/icaci55529.2022.9837693","DOIUrl":"https://doi.org/10.1109/icaci55529.2022.9837693","url":null,"abstract":"In this paper, a locally active hyperbolic memristor is proposed which characteristics are analyzed by numerical analysis. The HR-FN neuron model which bidirectional coupled by locally active hyperbolic memristor is constructed. Theoretical analysis and numerical simulation show that the bidirectional coupled neuron model has multiple stability under the influence of locally active memristor. Finally, the equivalent circuit of bidirectional coupled neural network is designed, and its the numerical analysis is verified by Multisim circuit simulation.","PeriodicalId":412347,"journal":{"name":"2022 14th International Conference on Advanced Computational Intelligence (ICACI)","volume":"586 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":"115103923","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/icaci55529.2022.9837571
Yingying Dong, Jianmin Wang
This paper studies the globally exponential stability of the equilibrium point for uncertain memristor-based recurrent neural networks (MRNN) with unbounded time-varying delay. The MRNN in this paper is the extension of classical MRNN since the uncertain factors and unbounded time-varying delay are considered. Under some assumptions for the MRNN, the equilibrium point of MRNN is proved to be globally exponentially stable by the Lyapunov method. A numerical experiment is performed to show the proposed result.
{"title":"Globally Exponential Stability of Uncertain Memristor-based Recurrent Neural Networks with Unbounded Time-varying Delays","authors":"Yingying Dong, Jianmin Wang","doi":"10.1109/icaci55529.2022.9837571","DOIUrl":"https://doi.org/10.1109/icaci55529.2022.9837571","url":null,"abstract":"This paper studies the globally exponential stability of the equilibrium point for uncertain memristor-based recurrent neural networks (MRNN) with unbounded time-varying delay. The MRNN in this paper is the extension of classical MRNN since the uncertain factors and unbounded time-varying delay are considered. Under some assumptions for the MRNN, the equilibrium point of MRNN is proved to be globally exponentially stable by the Lyapunov method. A numerical experiment is performed to show the proposed result.","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":"131655053","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 this paper, a novel two-layer memristive spiking neural network (MSNN) with spatio-temporal backpropagation (STBP) algorithm is proposed. The MSNN is composed of a memristive crossbar array, ten leaky integrate-and-fire (LIF) neurons and a winner-take-all (WTA) module. The memristive crossbar array with one memristor (1M) synapse performs the multiplication and accumulation without additional storage units. LIF neurons accumulate input current and fire spikes. WTA module ensures that only one neuron fires for one input pattern. The MSNN consumes a little power because there are no amplifiers or digital CMOS elements in the circuit. In order to train the memristive conductance, the LIF model is discretized and the gradients are propagated by the STBP algorithm. Furthermore, we perform a 6 × 5 black and white image classification based on the MSNN in PSPICE. Results verify that the MSNN realizes high recognition rates even under severe random noise and stuck-at faults
{"title":"A Novel Two-Layer Memristive Spiking Neural Network with Spatio-Temporal Backpropagation","authors":"Yaozhong Zhang, Mingxuan Jiang, Xiaoping Wang, Zhigang Zeng","doi":"10.1109/icaci55529.2022.9837606","DOIUrl":"https://doi.org/10.1109/icaci55529.2022.9837606","url":null,"abstract":"In this paper, a novel two-layer memristive spiking neural network (MSNN) with spatio-temporal backpropagation (STBP) algorithm is proposed. The MSNN is composed of a memristive crossbar array, ten leaky integrate-and-fire (LIF) neurons and a winner-take-all (WTA) module. The memristive crossbar array with one memristor (1M) synapse performs the multiplication and accumulation without additional storage units. LIF neurons accumulate input current and fire spikes. WTA module ensures that only one neuron fires for one input pattern. The MSNN consumes a little power because there are no amplifiers or digital CMOS elements in the circuit. In order to train the memristive conductance, the LIF model is discretized and the gradients are propagated by the STBP algorithm. Furthermore, we perform a 6 × 5 black and white image classification based on the MSNN in PSPICE. Results verify that the MSNN realizes high recognition rates even under severe random noise and stuck-at faults","PeriodicalId":412347,"journal":{"name":"2022 14th International Conference on Advanced Computational Intelligence (ICACI)","volume":"43 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":"121748120","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：