Pub Date : 2022-12-09DOI: 10.1109/ACAIT56212.2022.10138000
G. Luca, Yinong Chen
A major area of research in the field of quantum machine learning is the analysis of the loss landscape, particularly of variational quantum algorithms. These works often provide bounds and generalizations for various ansatzes and quantum embedding strategies. These analyses include approaches such as the Hessian and Fisher information matrices as well as generalized trigonometric polynomials. However, many such reviews often rely on a rotational encoding in practice or focus on few different approaches. The goal of this work is to statistically analyze experimental results from a quantum machine learning model that employs various different quantum embedding approaches, including those covered in related work, as well as the effect of measurement basis on the model.
{"title":"Quantum Embeddings of Classical Data for Quantum Machine Learning","authors":"G. Luca, Yinong Chen","doi":"10.1109/ACAIT56212.2022.10138000","DOIUrl":"https://doi.org/10.1109/ACAIT56212.2022.10138000","url":null,"abstract":"A major area of research in the field of quantum machine learning is the analysis of the loss landscape, particularly of variational quantum algorithms. These works often provide bounds and generalizations for various ansatzes and quantum embedding strategies. These analyses include approaches such as the Hessian and Fisher information matrices as well as generalized trigonometric polynomials. However, many such reviews often rely on a rotational encoding in practice or focus on few different approaches. The goal of this work is to statistically analyze experimental results from a quantum machine learning model that employs various different quantum embedding approaches, including those covered in related work, as well as the effect of measurement basis on the model.","PeriodicalId":398228,"journal":{"name":"2022 6th Asian Conference on Artificial Intelligence Technology (ACAIT)","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131115163","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 : 2022-12-09DOI: 10.1109/ACAIT56212.2022.10137862
Haitao Yu, Juntao Zeng, Xiaofeng Xie
Osteoporosis is a global skeletal disease which will seriously affect the human life. The early diagnosis of osteoporosis by using bone mineral density (BMD) examination can help to decrease the probability of osteoporosis. In the development of computer aided diagnosis, the calculation of BMD can be achieved by deep learning model in CT, without using the specially measuring devices. In this paper, we used a 3D-Unet model to segment the cortical and cancellous bone in the spine and perform quantitative analysis. After that, the three-dimensional visualization of cortical and cancellous bone was reconstructed, and the BMD value and other information were calculated to help doctors to predict the risk of osteoporosis. The expeirmental result shown that the proposed method achieve high performance in segementation and quantization.
{"title":"Deep Learning Model Research for Cortical Bone Separation in Chest CT Spine Imaging","authors":"Haitao Yu, Juntao Zeng, Xiaofeng Xie","doi":"10.1109/ACAIT56212.2022.10137862","DOIUrl":"https://doi.org/10.1109/ACAIT56212.2022.10137862","url":null,"abstract":"Osteoporosis is a global skeletal disease which will seriously affect the human life. The early diagnosis of osteoporosis by using bone mineral density (BMD) examination can help to decrease the probability of osteoporosis. In the development of computer aided diagnosis, the calculation of BMD can be achieved by deep learning model in CT, without using the specially measuring devices. In this paper, we used a 3D-Unet model to segment the cortical and cancellous bone in the spine and perform quantitative analysis. After that, the three-dimensional visualization of cortical and cancellous bone was reconstructed, and the BMD value and other information were calculated to help doctors to predict the risk of osteoporosis. The expeirmental result shown that the proposed method achieve high performance in segementation and quantization.","PeriodicalId":398228,"journal":{"name":"2022 6th Asian Conference on Artificial Intelligence Technology (ACAIT)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131206141","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 : 2022-12-09DOI: 10.1109/ACAIT56212.2022.10138001
Kaiyue Sun, Qiaoming Li, Wenlong Wang, P. Zhang, Zhantu Li, Xingnan Zhao, Zeqi Li
Loess landslide geological disasters are widely distributed in Northwest China, but there are few relevant attention and researches. Landslide recognition can provide information help for landslide disaster management and risk management. Previous works of landslide recognition of remote sensing images based on deep learning, due to the lack of high resolution multi-source datasets, the boundary of landslide recognition is missing and not obvious and the identification accuracy is not ideal. In this work, a multi-scale dense feature fusion loess landslide recognition network (MDFF) was proposed and an open dataset of loess landslide samples (MSLLD) based on GF-2 images and DEM was constructed, which has spectral and topographic information. The MDFF network retains different levels of features by means of dense connection mechanism to make up for the loss of detailed features, the dense connected dilated convolution layer is introduced into the network to capture the different scale features of landslide images, expand the receptive field and avoid convolution degradation. When testing different networks on MSLLD, the proposed network achieves the most advanced performance, mIoU and F1-score were 82.31 % and 84.59% respectively, indicating that the proposed network can effectively recognize landslides, which is of great value for the investigation and analysis of loess landslide disasters.
{"title":"Multi-Scale Dense Feature Fusion Based Loess Landslide Recognition","authors":"Kaiyue Sun, Qiaoming Li, Wenlong Wang, P. Zhang, Zhantu Li, Xingnan Zhao, Zeqi Li","doi":"10.1109/ACAIT56212.2022.10138001","DOIUrl":"https://doi.org/10.1109/ACAIT56212.2022.10138001","url":null,"abstract":"Loess landslide geological disasters are widely distributed in Northwest China, but there are few relevant attention and researches. Landslide recognition can provide information help for landslide disaster management and risk management. Previous works of landslide recognition of remote sensing images based on deep learning, due to the lack of high resolution multi-source datasets, the boundary of landslide recognition is missing and not obvious and the identification accuracy is not ideal. In this work, a multi-scale dense feature fusion loess landslide recognition network (MDFF) was proposed and an open dataset of loess landslide samples (MSLLD) based on GF-2 images and DEM was constructed, which has spectral and topographic information. The MDFF network retains different levels of features by means of dense connection mechanism to make up for the loss of detailed features, the dense connected dilated convolution layer is introduced into the network to capture the different scale features of landslide images, expand the receptive field and avoid convolution degradation. When testing different networks on MSLLD, the proposed network achieves the most advanced performance, mIoU and F1-score were 82.31 % and 84.59% respectively, indicating that the proposed network can effectively recognize landslides, which is of great value for the investigation and analysis of loess landslide disasters.","PeriodicalId":398228,"journal":{"name":"2022 6th Asian Conference on Artificial Intelligence Technology (ACAIT)","volume":"55 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133298458","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 : 2022-12-09DOI: 10.1109/ACAIT56212.2022.10137845
Xuefeng Liu, Guangjian Zhang
Grasping skills are the basic skills required by robots in many practical applications. Recent research on robotic grasping detection generally focuses on grasping poses similar to human grasping. However, this grasping pose is not suitable for all grasping scenarios in practical applications. Therefore, this paper uses a new inside-propped grasping pose to label a large number of images with inside-propped grasping potential. In this way, an inside-propped grasp dataset is completed. Based on this dataset, this paper constructs a generative deep neural network for the inside-propped grasping prediction. The experimental results show that the success rate of the inside-propped grasping prediction network is 65.59%, and the average prediction time is 82ms, which has achieved good results in accuracy and real-time performance.
{"title":"IPGD: A Dataset for Robotic Inside-Propped Grasp Detection","authors":"Xuefeng Liu, Guangjian Zhang","doi":"10.1109/ACAIT56212.2022.10137845","DOIUrl":"https://doi.org/10.1109/ACAIT56212.2022.10137845","url":null,"abstract":"Grasping skills are the basic skills required by robots in many practical applications. Recent research on robotic grasping detection generally focuses on grasping poses similar to human grasping. However, this grasping pose is not suitable for all grasping scenarios in practical applications. Therefore, this paper uses a new inside-propped grasping pose to label a large number of images with inside-propped grasping potential. In this way, an inside-propped grasp dataset is completed. Based on this dataset, this paper constructs a generative deep neural network for the inside-propped grasping prediction. The experimental results show that the success rate of the inside-propped grasping prediction network is 65.59%, and the average prediction time is 82ms, which has achieved good results in accuracy and real-time performance.","PeriodicalId":398228,"journal":{"name":"2022 6th Asian Conference on Artificial Intelligence Technology (ACAIT)","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134265270","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 : 2022-12-09DOI: 10.1109/ACAIT56212.2022.10137941
Baofu Fang, Shuai Zhou, Hao Wang
Most of the existing slam algorithms are designed based on the assumption of a static environment, this strong assumption limits the practical application of most slam systems. The main reason is that moving objects will cause feature mismatch in the pose estimation process, which in turn affects the accuracy of localization and mapping. In this paper, we propose a SLAM algorithm in a dynamic environment. First, we use the BlendMask network to detect potential moving objects to generate masks for dynamic objects. The geometrically constrained joint optical flow method is used to detect dynamic feature points. Secondly, aiming at the failure of semantic segmentation network segmentation, a missed detection compensation algorithm based on the invariance of adjacent frame speed is proposed. Finally, a keyframe selection strategy is proposed to construct a semantic octree graph containing only static objects. We evaluate our algorithm on TUM RGB-D and real scene datasets. The experimental results show that the algorithm has high accuracy and real-time performance.
{"title":"Semantic SLAM Based on Compensated Segmentation and Geometric Constraints in Dynamic Environments","authors":"Baofu Fang, Shuai Zhou, Hao Wang","doi":"10.1109/ACAIT56212.2022.10137941","DOIUrl":"https://doi.org/10.1109/ACAIT56212.2022.10137941","url":null,"abstract":"Most of the existing slam algorithms are designed based on the assumption of a static environment, this strong assumption limits the practical application of most slam systems. The main reason is that moving objects will cause feature mismatch in the pose estimation process, which in turn affects the accuracy of localization and mapping. In this paper, we propose a SLAM algorithm in a dynamic environment. First, we use the BlendMask network to detect potential moving objects to generate masks for dynamic objects. The geometrically constrained joint optical flow method is used to detect dynamic feature points. Secondly, aiming at the failure of semantic segmentation network segmentation, a missed detection compensation algorithm based on the invariance of adjacent frame speed is proposed. Finally, a keyframe selection strategy is proposed to construct a semantic octree graph containing only static objects. We evaluate our algorithm on TUM RGB-D and real scene datasets. The experimental results show that the algorithm has high accuracy and real-time performance.","PeriodicalId":398228,"journal":{"name":"2022 6th Asian Conference on Artificial Intelligence Technology (ACAIT)","volume":"45 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125332183","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 gradual expansion of the scale of”coal to electricity” users, the number of complaints about the heating effect, electricity safety and heating equipment safety guarantee in the clean heating operation process continues increasing. It is not possible to warn of possible emergencies in advance, and can be only remedied afterwards, completely in a passive response state. Therefore, rapid and accurate positioning of the key link is an urgent problem to be solved, but also the key to improve user satisfaction. Aimed at above problems, this paper established an automatic, information and intelligent electric heating risk warning mechanism. Based on the embedded feature selection algorithm and the DBSCAN adaptive clustering algorithm, a standardized vocabulary of customer appeals and complaint topics were constructed, combined with user historical electricity consumption data, and through the monitoring and matching of risk topics, an early warning model of customer electric heating abnormal risks was established. The model proposed in the article has strong practicability and provides strong support for lean management on the grid side, precise positioning of problems on the operation and maintenance side, government-side management decisionmaking and user satisfaction, and can promote safe, reliable and economical operation of the grid.
{"title":"Analysis and Research on Electric Heating Risk Early Warning Based on Embedded Feature Selection and DBSCAN Adaptive Clustering","authors":"Hui Xu, Lu Zhang, Longfei Ma, Xianglong Li, Siyue Lu, Shaokun Chen, Yifeng Ding, Wenbin Zhou","doi":"10.1109/ACAIT56212.2022.10137835","DOIUrl":"https://doi.org/10.1109/ACAIT56212.2022.10137835","url":null,"abstract":"With the gradual expansion of the scale of”coal to electricity” users, the number of complaints about the heating effect, electricity safety and heating equipment safety guarantee in the clean heating operation process continues increasing. It is not possible to warn of possible emergencies in advance, and can be only remedied afterwards, completely in a passive response state. Therefore, rapid and accurate positioning of the key link is an urgent problem to be solved, but also the key to improve user satisfaction. Aimed at above problems, this paper established an automatic, information and intelligent electric heating risk warning mechanism. Based on the embedded feature selection algorithm and the DBSCAN adaptive clustering algorithm, a standardized vocabulary of customer appeals and complaint topics were constructed, combined with user historical electricity consumption data, and through the monitoring and matching of risk topics, an early warning model of customer electric heating abnormal risks was established. The model proposed in the article has strong practicability and provides strong support for lean management on the grid side, precise positioning of problems on the operation and maintenance side, government-side management decisionmaking and user satisfaction, and can promote safe, reliable and economical operation of the grid.","PeriodicalId":398228,"journal":{"name":"2022 6th Asian Conference on Artificial Intelligence Technology (ACAIT)","volume":"14 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132522586","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 : 2022-12-09DOI: 10.1109/ACAIT56212.2022.10137854
Yaqi Chen, Dan Qu, Wenlin Zhang, Fen Yu, Haotong Zhang, Xukui Yang
Low-resource automatic speech recognition is a chal- lenging task. To solve this issue, multilingual meta-learning learns a better model initialization from many source language tasks, allowing for rapid adaption to the target language. However, due to the lack of limitations on multilingual pre-training, the shared semantic space of different languages is difficult to learn. In this work, we propose an adversarial meta-learning training approach to solve this problem. By using the adversarial auxiliary aim of language identification in the meta-learning algorithm, it will guide the model encoder to generate language-independent embedding features, which can improve model generalization. And we use Wasserstein distance and temporal normalization to optimize our adversarial training, making the training more stable and easier. The approach is evaluated on the IARPA BABEL. The results reveal that our approach only requires half as many meta learning training epochs to attain comparable multilingual pre-training performance. It also outperforms the meta learning in all target languages fine-tuning and achieves comparable performance in small data scales. Specially, it can reduce CER from 71% to 62% with fine-tuning 25% of Vietnamese data. Finally, we show why our approach is superior than others by using t-SNE.
{"title":"Adversarial Meta Learning Improves Low-Resource Speech Recognition","authors":"Yaqi Chen, Dan Qu, Wenlin Zhang, Fen Yu, Haotong Zhang, Xukui Yang","doi":"10.1109/ACAIT56212.2022.10137854","DOIUrl":"https://doi.org/10.1109/ACAIT56212.2022.10137854","url":null,"abstract":"Low-resource automatic speech recognition is a chal- lenging task. To solve this issue, multilingual meta-learning learns a better model initialization from many source language tasks, allowing for rapid adaption to the target language. However, due to the lack of limitations on multilingual pre-training, the shared semantic space of different languages is difficult to learn. In this work, we propose an adversarial meta-learning training approach to solve this problem. By using the adversarial auxiliary aim of language identification in the meta-learning algorithm, it will guide the model encoder to generate language-independent embedding features, which can improve model generalization. And we use Wasserstein distance and temporal normalization to optimize our adversarial training, making the training more stable and easier. The approach is evaluated on the IARPA BABEL. The results reveal that our approach only requires half as many meta learning training epochs to attain comparable multilingual pre-training performance. It also outperforms the meta learning in all target languages fine-tuning and achieves comparable performance in small data scales. Specially, it can reduce CER from 71% to 62% with fine-tuning 25% of Vietnamese data. Finally, we show why our approach is superior than others by using t-SNE.","PeriodicalId":398228,"journal":{"name":"2022 6th Asian Conference on Artificial Intelligence Technology (ACAIT)","volume":"37 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133881718","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 : 2022-12-09DOI: 10.1109/ACAIT56212.2022.10137999
Ruoyu Lou, Wu Yang, Yingjiang Li, Ling Lu
Aiming at the problem that the target tracking method of deep learning has a large number of model parameters and insufficient real-time performance, it is difficult to apply to mobile terminals or embedded devices with insufficient computing power. A lightweight hybrid attention-based twin network tracking algorithm is proposed. Firstly, based on MobileNetv3-Large network, group convolution and channel rearrangement are performed; then, in view of the problem that traditional attention mechanism only considers a single scope, this paper proposes a lightweight group-gated mixed attention (Group-gated mixed attention, GG); finally, GG is embedded in the Siamese network structure of this paper and the hierarchical feature fusion strategy is used to improve the tracking accuracy. Experiments show that the parameters of the proposed GG decrease by 26.2% compared with CBAM, decrease by 6.50% compared with SE, and increase Top-1 by 2.59% and 2.68% respectively; the experiments on the OTB100 and VOT2018 datasets demonstrate that the proposed algorithm is comparable to traditional tracking Compared with the algorithm, the accuracy and real-time performance have great advantages.
{"title":"Object Tracking Method Combined with Lightweight Hybrid Attention Siamese Network","authors":"Ruoyu Lou, Wu Yang, Yingjiang Li, Ling Lu","doi":"10.1109/ACAIT56212.2022.10137999","DOIUrl":"https://doi.org/10.1109/ACAIT56212.2022.10137999","url":null,"abstract":"Aiming at the problem that the target tracking method of deep learning has a large number of model parameters and insufficient real-time performance, it is difficult to apply to mobile terminals or embedded devices with insufficient computing power. A lightweight hybrid attention-based twin network tracking algorithm is proposed. Firstly, based on MobileNetv3-Large network, group convolution and channel rearrangement are performed; then, in view of the problem that traditional attention mechanism only considers a single scope, this paper proposes a lightweight group-gated mixed attention (Group-gated mixed attention, GG); finally, GG is embedded in the Siamese network structure of this paper and the hierarchical feature fusion strategy is used to improve the tracking accuracy. Experiments show that the parameters of the proposed GG decrease by 26.2% compared with CBAM, decrease by 6.50% compared with SE, and increase Top-1 by 2.59% and 2.68% respectively; the experiments on the OTB100 and VOT2018 datasets demonstrate that the proposed algorithm is comparable to traditional tracking Compared with the algorithm, the accuracy and real-time performance have great advantages.","PeriodicalId":398228,"journal":{"name":"2022 6th Asian Conference on Artificial Intelligence Technology (ACAIT)","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133565316","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 : 2022-12-09DOI: 10.1109/ACAIT56212.2022.10137951
Zhijian Xu
In order to improve the financial evaluation ability of pledged repo transactions, a method of identifying abnormal financial fluctuations of pledged repo transactions based on machine learning is proposed. Using the method of market risk identification, the pledge risk index system evaluation model for the financial evaluation of pledge type repo transactions is constructed. The balance of the capital flow channel of the pledge type repo financial system is controlled by using machine learning algorithm. Combined with machine learning to extract the abnormal fluctuation characteristics of the pledge type repo financial system, the fuzzy classification learning model of the data structure of the pledge type repo financial system is constructed. Spatial resampling method is used to reconstruct the abnormal financial volatility of pledge repurchase transactions and mining association rules. Clustering and matching the abnormal feature spectrum of the structural data of the financial system of pledge repurchase transactions by using machine learning algorithms. The model adopts the evaluation method of fluctuation synergy parameter. An adaptive learning algorithm is used to identify the abnormal financial fluctuations of pledge repurchase transactions. The simulation results show that this method has good clustering characteristics in identifying the abnormal financial fluctuations of pledge type repo transactions, effectively reducing the capital loss of the financial system structure of pledge type repo transactions, and improving the risk management ability.
{"title":"Research on Identification of Financial Abnormal Fluctuations in Pledged Repurchase Transactions Based on Machine Learning","authors":"Zhijian Xu","doi":"10.1109/ACAIT56212.2022.10137951","DOIUrl":"https://doi.org/10.1109/ACAIT56212.2022.10137951","url":null,"abstract":"In order to improve the financial evaluation ability of pledged repo transactions, a method of identifying abnormal financial fluctuations of pledged repo transactions based on machine learning is proposed. Using the method of market risk identification, the pledge risk index system evaluation model for the financial evaluation of pledge type repo transactions is constructed. The balance of the capital flow channel of the pledge type repo financial system is controlled by using machine learning algorithm. Combined with machine learning to extract the abnormal fluctuation characteristics of the pledge type repo financial system, the fuzzy classification learning model of the data structure of the pledge type repo financial system is constructed. Spatial resampling method is used to reconstruct the abnormal financial volatility of pledge repurchase transactions and mining association rules. Clustering and matching the abnormal feature spectrum of the structural data of the financial system of pledge repurchase transactions by using machine learning algorithms. The model adopts the evaluation method of fluctuation synergy parameter. An adaptive learning algorithm is used to identify the abnormal financial fluctuations of pledge repurchase transactions. The simulation results show that this method has good clustering characteristics in identifying the abnormal financial fluctuations of pledge type repo transactions, effectively reducing the capital loss of the financial system structure of pledge type repo transactions, and improving the risk management ability.","PeriodicalId":398228,"journal":{"name":"2022 6th Asian Conference on Artificial Intelligence Technology (ACAIT)","volume":"29 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117137966","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 : 2022-12-09DOI: 10.1109/ACAIT56212.2022.10137929
Dengyi Xiao
In order to correct business English translation errors, this paper puts forward a method of business English translation error correction based on convolutional neural network and English pronunciation feature recognition. The blind convolution network spectrum parameter detection method is used to detect the pronunciation spectrum features of business English translation, and the scalar time series of pronunciation output audio parameter sequence and translated text semantic feature sequence are established. Combined with the noise intensity detection and signal scale decomposition method of business English translation pronunciation audio time series, the detailed signal energy parameters of business English translation pronunciation audio time series are extracted, and the convolution neural network classification method is used to classify the features. The interference component of single audio feature sequence of English translation pronunciation is removed by high-frequency wavelet threshold detection, and the modulation and demodulation of single audio feature sequence of English translation pronunciation are realized by using translation dictionary set and semantic context matching. The spectral analysis and error correction model of business English translation pronunciation audio time series is established, and the output stability of business English translation pronunciation audio time series is detected by threshold detection on each scale. According to the difference between output signal and pronunciation standard signal, the accuracy of English translator is detected and identified. The simulation results show that the accuracy of business English translation error correction with this method is high, the detection performance is good, and the output accuracy of English translators is improved.
{"title":"Error Correction Method of Business English Translation Based on Convolutional Neural Network","authors":"Dengyi Xiao","doi":"10.1109/ACAIT56212.2022.10137929","DOIUrl":"https://doi.org/10.1109/ACAIT56212.2022.10137929","url":null,"abstract":"In order to correct business English translation errors, this paper puts forward a method of business English translation error correction based on convolutional neural network and English pronunciation feature recognition. The blind convolution network spectrum parameter detection method is used to detect the pronunciation spectrum features of business English translation, and the scalar time series of pronunciation output audio parameter sequence and translated text semantic feature sequence are established. Combined with the noise intensity detection and signal scale decomposition method of business English translation pronunciation audio time series, the detailed signal energy parameters of business English translation pronunciation audio time series are extracted, and the convolution neural network classification method is used to classify the features. The interference component of single audio feature sequence of English translation pronunciation is removed by high-frequency wavelet threshold detection, and the modulation and demodulation of single audio feature sequence of English translation pronunciation are realized by using translation dictionary set and semantic context matching. The spectral analysis and error correction model of business English translation pronunciation audio time series is established, and the output stability of business English translation pronunciation audio time series is detected by threshold detection on each scale. According to the difference between output signal and pronunciation standard signal, the accuracy of English translator is detected and identified. The simulation results show that the accuracy of business English translation error correction with this method is high, the detection performance is good, and the output accuracy of English translators is improved.","PeriodicalId":398228,"journal":{"name":"2022 6th Asian Conference on Artificial Intelligence Technology (ACAIT)","volume":"75 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114996089","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：