One-shot face recognition is a challenging problem which requires recognizing novel identities from only one seen face image. One-shot classes are simply neglected because of the lack of training samples. Therefore, these classes contribute less to the improvement of face recognition performance. The main goal of one-shot face recognition task is to use the novel face samples to enhance the ability of network not only in close-set classify, but also in open-set face verification. In this paper, Base data and Novel data is trained separately with two classifiers to reduce the impact of data imbalance. We propose Confidence Constrain Loss to train classifiers in parallel and get better classifiers fusion in the test phase. Besides, we use data augmentation with 3D face reconstruction to obtain a variety of oneshot set's training samples. Thus, our method can effectively increase the recognition accuracy in the novel set without reducing recognition accuracy in base set. Experiments on MS-celeb-1M low-shot dataset demonstrate that our method achieve state-of-the-art which has 98.90% coverage at precision=99% without using external data.
{"title":"One-Shot Face Recognition Based on Multiple Classifiers Training","authors":"Vuliem Khong, Ziyu Zeng, Lu Fang, Shengjin Wang","doi":"10.1145/3457682.3457748","DOIUrl":"https://doi.org/10.1145/3457682.3457748","url":null,"abstract":"One-shot face recognition is a challenging problem which requires recognizing novel identities from only one seen face image. One-shot classes are simply neglected because of the lack of training samples. Therefore, these classes contribute less to the improvement of face recognition performance. The main goal of one-shot face recognition task is to use the novel face samples to enhance the ability of network not only in close-set classify, but also in open-set face verification. In this paper, Base data and Novel data is trained separately with two classifiers to reduce the impact of data imbalance. We propose Confidence Constrain Loss to train classifiers in parallel and get better classifiers fusion in the test phase. Besides, we use data augmentation with 3D face reconstruction to obtain a variety of oneshot set's training samples. Thus, our method can effectively increase the recognition accuracy in the novel set without reducing recognition accuracy in base set. Experiments on MS-celeb-1M low-shot dataset demonstrate that our method achieve state-of-the-art which has 98.90% coverage at precision=99% without using external data.","PeriodicalId":142045,"journal":{"name":"2021 13th International Conference on Machine Learning and Computing","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-02-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133923097","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}
The cocktail party effect is a fundamental problem in sound source separation, and many researchers have worked to solve this problem. In recent years, the most popular algorithms to solve the problem of sound source separation are Support Vector Machine (SVM), Gaussian Mixture Model (GMM), non-negative matrix factorization (NMF), and Variational Autoencoder (VAE). Especially VAE model showed excellent ability in dealing with the problem of sound separation. In this paper, the β-VAE model, combined with a weakly supervised classification proposed by Karamatlı et al., was first reproduced. Since Karamatlı's experiment only completed the connection between sound and words, in order to learn more information about the speaker, this model is used to learn a mapping between sounds and individual speakers and a mapping between sounds and gender. It turns out that the separation results could be obtained by retraining the model after the establishment of the new 'male' and 'female' labels. his result lays a foundation for the future study of the mapping between individuals and words. When the tag is specific to an individual, more data is needed to support this experiment, and the more data available for training, the better result the model will get.
{"title":"Research on Deep Sound Source Separation","authors":"Yunuo Yang, Honghui Li","doi":"10.1145/3457682.3457741","DOIUrl":"https://doi.org/10.1145/3457682.3457741","url":null,"abstract":"The cocktail party effect is a fundamental problem in sound source separation, and many researchers have worked to solve this problem. In recent years, the most popular algorithms to solve the problem of sound source separation are Support Vector Machine (SVM), Gaussian Mixture Model (GMM), non-negative matrix factorization (NMF), and Variational Autoencoder (VAE). Especially VAE model showed excellent ability in dealing with the problem of sound separation. In this paper, the β-VAE model, combined with a weakly supervised classification proposed by Karamatlı et al., was first reproduced. Since Karamatlı's experiment only completed the connection between sound and words, in order to learn more information about the speaker, this model is used to learn a mapping between sounds and individual speakers and a mapping between sounds and gender. It turns out that the separation results could be obtained by retraining the model after the establishment of the new 'male' and 'female' labels. his result lays a foundation for the future study of the mapping between individuals and words. When the tag is specific to an individual, more data is needed to support this experiment, and the more data available for training, the better result the model will get.","PeriodicalId":142045,"journal":{"name":"2021 13th International Conference on Machine Learning and Computing","volume":"126 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-02-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124209903","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This paper discusses the application of a special data augmentation approach for end-to-end phone recognition system on the Deep Neural Networks. The system improves the performance of phone recognition and alleviates overfitting during training. Also, it offers a solution to the problem of few public datasets annotated at the phone level. And we propose the CNN-CTC structure as a baseline model. The model is based on Convolutional Neural Networks (CNNs) and Connectionist Temporal Classification (CTC) objective function. Which is an end-to-end structure, and there is no need to force alignment each frame of audio. The SpecAugment approach directly processes the feature of audio, such as the log Mel-spectrogram. In our experiment, the Spec-CNN-CTC system achieves a phone error rate of 16.11% on TIMIT corpus with no prior linguistic information. Which is outperforming the previous work Acoustic-State-Transition Model (ASTM) by 27.63%, the DNN-HMM with MFCC + IFCC features by 16.8%, the RNN-CRF model by 17.3% and the DBM-DNN model by 22.62%.
{"title":"A Novel Spec-CNN-CTC Model for End-to-End Speech Recognition","authors":"Jing Xue, Jun Zhang","doi":"10.1145/3457682.3457703","DOIUrl":"https://doi.org/10.1145/3457682.3457703","url":null,"abstract":"This paper discusses the application of a special data augmentation approach for end-to-end phone recognition system on the Deep Neural Networks. The system improves the performance of phone recognition and alleviates overfitting during training. Also, it offers a solution to the problem of few public datasets annotated at the phone level. And we propose the CNN-CTC structure as a baseline model. The model is based on Convolutional Neural Networks (CNNs) and Connectionist Temporal Classification (CTC) objective function. Which is an end-to-end structure, and there is no need to force alignment each frame of audio. The SpecAugment approach directly processes the feature of audio, such as the log Mel-spectrogram. In our experiment, the Spec-CNN-CTC system achieves a phone error rate of 16.11% on TIMIT corpus with no prior linguistic information. Which is outperforming the previous work Acoustic-State-Transition Model (ASTM) by 27.63%, the DNN-HMM with MFCC + IFCC features by 16.8%, the RNN-CRF model by 17.3% and the DBM-DNN model by 22.62%.","PeriodicalId":142045,"journal":{"name":"2021 13th International Conference on Machine Learning and Computing","volume":"17 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-02-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121212512","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 recent years, the increasing influence of machine learning in different industries had inspired many traders to benefit from it in the world of finance, stock trading is one of the most important activities. Predicting the direction of stock prices is a widely studied subject in many fields including trading, finance, statistics and computer science. The main concern for Investors is to maximize their profit if they determine when to buy/sell an investment they apply Analytical methods that makes use of different sources ranging from news to price data, all aiming at predicting the company's future stock price ML applications have presented investors with something new. A combination of technologies that could entirely reshape the way they make investment decisions. The purpose of this thesis is to leverage the aggregation of technical, fundamental, and sentiment analysis with stacked machine learning models capable of predicting profitable actions to be executed.
{"title":"Ensemble Learning in Stock Market Prediction","authors":"Hassan Ezzeddine, Roger Achkar","doi":"10.1145/3457682.3457727","DOIUrl":"https://doi.org/10.1145/3457682.3457727","url":null,"abstract":"In recent years, the increasing influence of machine learning in different industries had inspired many traders to benefit from it in the world of finance, stock trading is one of the most important activities. Predicting the direction of stock prices is a widely studied subject in many fields including trading, finance, statistics and computer science. The main concern for Investors is to maximize their profit if they determine when to buy/sell an investment they apply Analytical methods that makes use of different sources ranging from news to price data, all aiming at predicting the company's future stock price ML applications have presented investors with something new. A combination of technologies that could entirely reshape the way they make investment decisions. The purpose of this thesis is to leverage the aggregation of technical, fundamental, and sentiment analysis with stacked machine learning models capable of predicting profitable actions to be executed.","PeriodicalId":142045,"journal":{"name":"2021 13th International Conference on Machine Learning and Computing","volume":"39 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-02-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129293304","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}
The probability hypothesis density filter with linear Gaussian jump Markov system multi-target models is an attractive approach to tracking multiple maneuvering targets in the presence of data association uncertainty, clutter, noise, and detection uncertainty. However, these models are not precise enough to describe moving targets on road networks in ground target tracking scenario. In this paper, the road map information is integrated into the jump Markov system Gaussian mixture probability hypothesis density (JMS-GM-PHD) filter, and a road-constraint JMS-GM-PHD filter for ground target tracking is proposed. In addition, we then derive the recursive equation of the proposed filter. Simulation results show that the proposed road-constrained JMS-GM-PHD filter is effective in tracking ground moving targets.
{"title":"Tracking Ground Targets with Road Constraints Using a JMS-GM-PHD Filter","authors":"Jihong Zheng, He He, Longteng Cong","doi":"10.1145/3457682.3457768","DOIUrl":"https://doi.org/10.1145/3457682.3457768","url":null,"abstract":"The probability hypothesis density filter with linear Gaussian jump Markov system multi-target models is an attractive approach to tracking multiple maneuvering targets in the presence of data association uncertainty, clutter, noise, and detection uncertainty. However, these models are not precise enough to describe moving targets on road networks in ground target tracking scenario. In this paper, the road map information is integrated into the jump Markov system Gaussian mixture probability hypothesis density (JMS-GM-PHD) filter, and a road-constraint JMS-GM-PHD filter for ground target tracking is proposed. In addition, we then derive the recursive equation of the proposed filter. Simulation results show that the proposed road-constrained JMS-GM-PHD filter is effective in tracking ground moving targets.","PeriodicalId":142045,"journal":{"name":"2021 13th International Conference on Machine Learning and Computing","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-02-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116347745","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}
The fault logs record the fault information generated during the operation process of industrial robots. It contains a large amount of fault knowledge and solution information. It is necessary to extract this information and build the fault diagnosis knowledge graph of industrial robots, which can support remote fault diagnosis of industrial robots without human help. At present, the research of fault diagnosis knowledge graph is still relatively scarce. In this paper, we propose a method of named entity recognition for extracting the knowledge of industrial robot fault diagnosis. The contribution of our paper is to establish the fault field dataset Fault-Data, propose the ontology concept of the fault diagnosis field, and obtain a good field recognition effect through the verification of the entity recognition model of fault diagnosis. Experimental results show that the F value of named entity recognition reaches 91.99%, which provides a certain reference significance for subsequent knowledge extraction and knowledge graph construction.
{"title":"Corpus Construction and Entity Recognition for the Field of Industrial Robot Fault Diagnosis","authors":"Jiale Zhou, Tao Wang, Jianfeng Deng","doi":"10.1145/3457682.3457745","DOIUrl":"https://doi.org/10.1145/3457682.3457745","url":null,"abstract":"The fault logs record the fault information generated during the operation process of industrial robots. It contains a large amount of fault knowledge and solution information. It is necessary to extract this information and build the fault diagnosis knowledge graph of industrial robots, which can support remote fault diagnosis of industrial robots without human help. At present, the research of fault diagnosis knowledge graph is still relatively scarce. In this paper, we propose a method of named entity recognition for extracting the knowledge of industrial robot fault diagnosis. The contribution of our paper is to establish the fault field dataset Fault-Data, propose the ontology concept of the fault diagnosis field, and obtain a good field recognition effect through the verification of the entity recognition model of fault diagnosis. Experimental results show that the F value of named entity recognition reaches 91.99%, which provides a certain reference significance for subsequent knowledge extraction and knowledge graph construction.","PeriodicalId":142045,"journal":{"name":"2021 13th International Conference on Machine Learning and Computing","volume":"67 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-02-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114812374","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}
WeiCai Niu, Quan Chen, Weiwen Zhang, Jianwen Ma, Zhongqiang Hu
Joint extraction of entities and relations is critical for many tasks of Natural Language Processing (NLP), which aims to extract all triplets in the text. However, the huge challenge is that a sentence usually contains overlapping triplets. In this paper, we propose a joint extraction framework named GCN2-NAA based on a two-stage Graph Convolutional Neural networks (GCN) and Node-Aware Attention mechanism. We obtain multi-granularity representations and regional features of words by stacking multiple feature encoders and 1st-phase GCN. Besides, the node-aware attention mechanism and 2nd-phase GCN to capture the soft attention correlation matrix between all words in each relation type. Based on the constructed soft attention correlation matrix, we utilize GCN to further obtain the interaction between entities, relations, and triplets. Experiment results show that GCN2-NAA outperforms baseline models by 6.5% and 11.4% in terms of F1 score on NYT and WebNLG datasets, respectively.
{"title":"GCN2-NAA: Two-stage Graph Convolutional Networks with Node-Aware Attention for Joint Entity and Relation Extraction","authors":"WeiCai Niu, Quan Chen, Weiwen Zhang, Jianwen Ma, Zhongqiang Hu","doi":"10.1145/3457682.3457765","DOIUrl":"https://doi.org/10.1145/3457682.3457765","url":null,"abstract":"Joint extraction of entities and relations is critical for many tasks of Natural Language Processing (NLP), which aims to extract all triplets in the text. However, the huge challenge is that a sentence usually contains overlapping triplets. In this paper, we propose a joint extraction framework named GCN2-NAA based on a two-stage Graph Convolutional Neural networks (GCN) and Node-Aware Attention mechanism. We obtain multi-granularity representations and regional features of words by stacking multiple feature encoders and 1st-phase GCN. Besides, the node-aware attention mechanism and 2nd-phase GCN to capture the soft attention correlation matrix between all words in each relation type. Based on the constructed soft attention correlation matrix, we utilize GCN to further obtain the interaction between entities, relations, and triplets. Experiment results show that GCN2-NAA outperforms baseline models by 6.5% and 11.4% in terms of F1 score on NYT and WebNLG datasets, respectively.","PeriodicalId":142045,"journal":{"name":"2021 13th International Conference on Machine Learning and Computing","volume":"45 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-02-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114822200","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}
The genotype imputation is an important topic in the field of genomics. Many genome analyses require data without missing values, which requires to impute the missing data. In recent years, deep learning has become hot, and it is more suitable for text sequence type problems, which may fit with the genotype imputation problem. Based on the recurrent neural network and convolutional neural network in deep learning, our study proposes and constructs five model combinations, imputes and compares the results under different missing rate scenarios. And on the basis of the basic model, a higher imputation accuracy is obtained by tuning the model hyperparameters. The results indicated that on all the data sets with various levels of missing rates, the CNN1D-RNNM with tuned hyperparameters well has obtained the best results. The combination of a one-dimensional convolutional neural network and a recurrent neural network with tuned hyperparameters can beat a single convolutional network or a recurrent network at various levels of missing rates. This research provides new solutions for genotype imputation by using the deep learning to build complex neural networks.
{"title":"Factors Affecting Accuracy of Genotype Imputation Using Neural Networks in Deep Learning","authors":"Tianfeng Shi, Jing Peng","doi":"10.1145/3457682.3457688","DOIUrl":"https://doi.org/10.1145/3457682.3457688","url":null,"abstract":"The genotype imputation is an important topic in the field of genomics. Many genome analyses require data without missing values, which requires to impute the missing data. In recent years, deep learning has become hot, and it is more suitable for text sequence type problems, which may fit with the genotype imputation problem. Based on the recurrent neural network and convolutional neural network in deep learning, our study proposes and constructs five model combinations, imputes and compares the results under different missing rate scenarios. And on the basis of the basic model, a higher imputation accuracy is obtained by tuning the model hyperparameters. The results indicated that on all the data sets with various levels of missing rates, the CNN1D-RNNM with tuned hyperparameters well has obtained the best results. The combination of a one-dimensional convolutional neural network and a recurrent neural network with tuned hyperparameters can beat a single convolutional network or a recurrent network at various levels of missing rates. This research provides new solutions for genotype imputation by using the deep learning to build complex neural networks.","PeriodicalId":142045,"journal":{"name":"2021 13th International Conference on Machine Learning and Computing","volume":"21 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-02-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127045806","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}
Anomaly detection is an significant problem in machine learning and has been well-studied in a wide range of applications. To model complex and high dimensional data distributions, existing methods, trained with an auto-encoder architecture either directly or indirectly, usually attempt to produce higher reconstruction error for anomalies than normal samples. However, lacking constrains on the latent representation of input data results in an unexpected performance that anomalies can be reconstructed well too, leading to the “missed alarm”. In this work, we propose to reconstruct input data with typical patterns of normal data learned through adversarial networks. Our approach, called MNGAN, which employs an encoder-decoder-encoder architecture with a memory network, learns to memorize prototypical patterns of normal data and simultaneously preserve details of data style for better reconstruction. In test phase, given a input data, the model will reconstruct it with the most relevant memory item, which indicates one normal pattern. Thus, reconstructions of anomalous data are similar to normal samples, resulting in effective detection for anomalies due to the high reconstruction error. Experiments over several benchmark datasets, from varying domains, shows that our proposed method outperforms previous state-of-the-art anomaly detection approaches.
{"title":"MNGAN: Detecting Anomalies with Memorized Normal Patterns","authors":"Zijian Huang, Changqing Xu","doi":"10.1145/3457682.3457764","DOIUrl":"https://doi.org/10.1145/3457682.3457764","url":null,"abstract":"Anomaly detection is an significant problem in machine learning and has been well-studied in a wide range of applications. To model complex and high dimensional data distributions, existing methods, trained with an auto-encoder architecture either directly or indirectly, usually attempt to produce higher reconstruction error for anomalies than normal samples. However, lacking constrains on the latent representation of input data results in an unexpected performance that anomalies can be reconstructed well too, leading to the “missed alarm”. In this work, we propose to reconstruct input data with typical patterns of normal data learned through adversarial networks. Our approach, called MNGAN, which employs an encoder-decoder-encoder architecture with a memory network, learns to memorize prototypical patterns of normal data and simultaneously preserve details of data style for better reconstruction. In test phase, given a input data, the model will reconstruct it with the most relevant memory item, which indicates one normal pattern. Thus, reconstructions of anomalous data are similar to normal samples, resulting in effective detection for anomalies due to the high reconstruction error. Experiments over several benchmark datasets, from varying domains, shows that our proposed method outperforms previous state-of-the-art anomaly detection approaches.","PeriodicalId":142045,"journal":{"name":"2021 13th International Conference on Machine Learning and Computing","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-02-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129916706","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}
Neural network has strong nonlinear data characterization ability and solves many complex problems successfully. Trajectory estimation and prediction is time series forecasting but different from convention problems such as time video analysis. A method based on neural network and error self-correction technology achieving trajectory estimation and prediction is proposed in this paper. The method needs neural network without additional filtering algorithm, so the maneuver models and noise characteristics are not needed. According to the information of the previous moments before the investigated time, the information of the next moment or a specified time later can be obtained. For tracking a simple maneuvering target model with unknown parameters and noise characteristics, numerical simulation results show that FNN achieves filtering and it achieves a higher prediction accuracy than the Least Squares filtering. For tracking a complex maneuvering target model with strong nonlinearity, RNN combining with FNN is employed. For the measurement error with D standard deviation 2m, azimuth angle and the altitude angle measurement errors standard deviation with 2mil, the angle predicting error standard deviation is less than 1.3mil, which shows RNN combing with error self-correction technology has high accuracy. It meets the technical requirements for maneuvering target tracking as well as various similar applications.
{"title":"Maneuvering Target Tracking Based on Neural Network and Error Self-correction Technology","authors":"Lisi Chen, Changcheng Wang, Jiale Huang","doi":"10.1145/3457682.3457708","DOIUrl":"https://doi.org/10.1145/3457682.3457708","url":null,"abstract":"Neural network has strong nonlinear data characterization ability and solves many complex problems successfully. Trajectory estimation and prediction is time series forecasting but different from convention problems such as time video analysis. A method based on neural network and error self-correction technology achieving trajectory estimation and prediction is proposed in this paper. The method needs neural network without additional filtering algorithm, so the maneuver models and noise characteristics are not needed. According to the information of the previous moments before the investigated time, the information of the next moment or a specified time later can be obtained. For tracking a simple maneuvering target model with unknown parameters and noise characteristics, numerical simulation results show that FNN achieves filtering and it achieves a higher prediction accuracy than the Least Squares filtering. For tracking a complex maneuvering target model with strong nonlinearity, RNN combining with FNN is employed. For the measurement error with D standard deviation 2m, azimuth angle and the altitude angle measurement errors standard deviation with 2mil, the angle predicting error standard deviation is less than 1.3mil, which shows RNN combing with error self-correction technology has high accuracy. It meets the technical requirements for maneuvering target tracking as well as various similar applications.","PeriodicalId":142045,"journal":{"name":"2021 13th International Conference on Machine Learning and Computing","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-02-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131071881","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：