Pub Date : 2019-07-01DOI: 10.1109/ICMLC48188.2019.8949286
Ting Yuan, Haihui Li, Hongya Zhao, Qianhua Cai, Han Liu, Xiaohui Hu
In recent years, targeted sentiment analysis has received great attention as a fine-grained sentiment analysis. Determining the sentiment polarity of a specific target in a sentence is the main task. This paper proposes a multi-channel convolutional neural network (MCL-CNN) for targeted sentiment classification. Our approach can not only parallelize over the words of a sentence but also extract local features effectively. Contexts and targets can be more comprehensively utilized by using part-of-speech information, semantic information and interactive information so that diverse features can be obtained. Finally, experimental results on the SemEval 2014 dataset demonstrate the effectiveness of this method.
{"title":"Multi-Channel Convolutional Neural Network for Targeted Sentiment Classification","authors":"Ting Yuan, Haihui Li, Hongya Zhao, Qianhua Cai, Han Liu, Xiaohui Hu","doi":"10.1109/ICMLC48188.2019.8949286","DOIUrl":"https://doi.org/10.1109/ICMLC48188.2019.8949286","url":null,"abstract":"In recent years, targeted sentiment analysis has received great attention as a fine-grained sentiment analysis. Determining the sentiment polarity of a specific target in a sentence is the main task. This paper proposes a multi-channel convolutional neural network (MCL-CNN) for targeted sentiment classification. Our approach can not only parallelize over the words of a sentence but also extract local features effectively. Contexts and targets can be more comprehensively utilized by using part-of-speech information, semantic information and interactive information so that diverse features can be obtained. Finally, experimental results on the SemEval 2014 dataset demonstrate the effectiveness of this method.","PeriodicalId":221349,"journal":{"name":"2019 International Conference on Machine Learning and Cybernetics (ICMLC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115536273","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}
Targeting older adults with high risks of dysphagia. This study introduced a wearable device for saliva detection and improvement in salivation. Said device could detect its users” long-term salivation situations and determine/issue the following when salivation occurred: whether the users salivated in their left or right cheeks and voice prompts reminding the users to swallow their saliva. In this way, excessive accumulation of saliva in the oral of the older adults is avoided., resulting in coughing and even pneumonia.
{"title":"A Wearable Device of Salivation Detection and Improvement for Elderly at High Risk for Dysphagia","authors":"Chien-Nan Lee, Meng-Hsuan Shih, Chia-Wei Chen, Ding-Jiun Tzeng, Chuan-Che Shih, Yiu-Tong Chu, Ling Cheng","doi":"10.1109/ICMLC48188.2019.8949257","DOIUrl":"https://doi.org/10.1109/ICMLC48188.2019.8949257","url":null,"abstract":"Targeting older adults with high risks of dysphagia. This study introduced a wearable device for saliva detection and improvement in salivation. Said device could detect its users” long-term salivation situations and determine/issue the following when salivation occurred: whether the users salivated in their left or right cheeks and voice prompts reminding the users to swallow their saliva. In this way, excessive accumulation of saliva in the oral of the older adults is avoided., resulting in coughing and even pneumonia.","PeriodicalId":221349,"journal":{"name":"2019 International Conference on Machine Learning and Cybernetics (ICMLC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116044572","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 : 2019-07-01DOI: 10.1109/ICMLC48188.2019.8949303
T. Anjiki, K. Matsubayashi, Shunji Maeda
Garbage power generation can play a key role because it can supply uninterrupted power among the various renewable energies. Generation can be performed using steam, which is generated by garbage incineration as an energy source. For an uninterrupted electric power supply, it is necessary to control the steam such that its flow becomes stable. Therefore, in this study, the possibility of the steam flow estimation using an infrared image of the furnace is examined
{"title":"Utilization of the Infrared Image Capturing Combustion State for Estimating the Steam Flow Aming to Stabilize Garbage Power Generation","authors":"T. Anjiki, K. Matsubayashi, Shunji Maeda","doi":"10.1109/ICMLC48188.2019.8949303","DOIUrl":"https://doi.org/10.1109/ICMLC48188.2019.8949303","url":null,"abstract":"Garbage power generation can play a key role because it can supply uninterrupted power among the various renewable energies. Generation can be performed using steam, which is generated by garbage incineration as an energy source. For an uninterrupted electric power supply, it is necessary to control the steam such that its flow becomes stable. Therefore, in this study, the possibility of the steam flow estimation using an infrared image of the furnace is examined","PeriodicalId":221349,"journal":{"name":"2019 International Conference on Machine Learning and Cybernetics (ICMLC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115114843","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 : 2019-07-01DOI: 10.1109/ICMLC48188.2019.8949185
J. Chai, Anming Li
Deep learning raises interests of research community as their overwhelming successes in information processing such specific tasks as video/speech recognition. In this paper, we provide a state-of-the-art analysis of deep learning with its applications in an important direction: natural language processing. We attempt to provide a clear and critical summarization for researchers and participators who are interested in incorporating the deep learning techniques in their specific domains.
{"title":"Deep Learning in Natural Language Processing: A State-of-the-Art Survey","authors":"J. Chai, Anming Li","doi":"10.1109/ICMLC48188.2019.8949185","DOIUrl":"https://doi.org/10.1109/ICMLC48188.2019.8949185","url":null,"abstract":"Deep learning raises interests of research community as their overwhelming successes in information processing such specific tasks as video/speech recognition. In this paper, we provide a state-of-the-art analysis of deep learning with its applications in an important direction: natural language processing. We attempt to provide a clear and critical summarization for researchers and participators who are interested in incorporating the deep learning techniques in their specific domains.","PeriodicalId":221349,"journal":{"name":"2019 International Conference on Machine Learning and Cybernetics (ICMLC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124970855","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 : 2019-07-01DOI: 10.1109/ICMLC48188.2019.8949225
Wei-Fu Kao, Chun-Fei Hsu
To be able to interact with humans, multiple rounds of statically stabilized mobile robots must have a low center of gravity and a large bottom area to avoid the robot tipping. This paper considers a dynamic spherical mobile robot (DSMR) system to overcome these mechanism limitations. To design the controller and system characteristic analysis, this paper proposes a motion controller design method which comprises a PD angle controller and a fuzzy position controller for the DSMR system. The PD angle controller can achieve the balance of movement control responses and the fuzzy position controller can achieve the favorable position control responses. Meanwhile, a steering controller is designed to obtain the robot rotation ability. Finally, the experimental results verifies that the proposed motion control system can achieve a good dynamic balance effect for the DSMR system even when there is an external force to push the robot.
{"title":"Motion Control Design for Dynamic Spherical Mobile Robot via Fuzzy Control Approach","authors":"Wei-Fu Kao, Chun-Fei Hsu","doi":"10.1109/ICMLC48188.2019.8949225","DOIUrl":"https://doi.org/10.1109/ICMLC48188.2019.8949225","url":null,"abstract":"To be able to interact with humans, multiple rounds of statically stabilized mobile robots must have a low center of gravity and a large bottom area to avoid the robot tipping. This paper considers a dynamic spherical mobile robot (DSMR) system to overcome these mechanism limitations. To design the controller and system characteristic analysis, this paper proposes a motion controller design method which comprises a PD angle controller and a fuzzy position controller for the DSMR system. The PD angle controller can achieve the balance of movement control responses and the fuzzy position controller can achieve the favorable position control responses. Meanwhile, a steering controller is designed to obtain the robot rotation ability. Finally, the experimental results verifies that the proposed motion control system can achieve a good dynamic balance effect for the DSMR system even when there is an external force to push the robot.","PeriodicalId":221349,"journal":{"name":"2019 International Conference on Machine Learning and Cybernetics (ICMLC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122137072","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 : 2019-07-01DOI: 10.1109/ICMLC48188.2019.8949251
Hui-Hui Chen, Chiao-Wen Kao, B. Hwang, Kuo-Chin Fan
This study proposes a novel approach using YOLO based deep learning network to help the teacher grading 3D modeling objects created by the learners automatically. The training dataset is the collections of rendering outputs from the teacher's 3D modeling object. The testing data is the rendering outputs of the learners' projects. The grading will rely on the testing results of recognition confidences. This is an initial study from draft inspiration by the deep learning network on object detections and recognitions. More applications and modifications are to be discussed, designed and examined in further studies.
{"title":"Recognizing and Grading 3D Modeling Objects Using YOLO Based Deep Learning Network","authors":"Hui-Hui Chen, Chiao-Wen Kao, B. Hwang, Kuo-Chin Fan","doi":"10.1109/ICMLC48188.2019.8949251","DOIUrl":"https://doi.org/10.1109/ICMLC48188.2019.8949251","url":null,"abstract":"This study proposes a novel approach using YOLO based deep learning network to help the teacher grading 3D modeling objects created by the learners automatically. The training dataset is the collections of rendering outputs from the teacher's 3D modeling object. The testing data is the rendering outputs of the learners' projects. The grading will rely on the testing results of recognition confidences. This is an initial study from draft inspiration by the deep learning network on object detections and recognitions. More applications and modifications are to be discussed, designed and examined in further studies.","PeriodicalId":221349,"journal":{"name":"2019 International Conference on Machine Learning and Cybernetics (ICMLC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125197446","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 : 2019-07-01DOI: 10.1109/ICMLC48188.2019.8949245
T. Murata, Takuya Harada, Manabu Ichika Wa, Yusuke Goto, Lee Hao, S. Date, M. Munetomo, Akiyoshi Sugiki
In this paper, we describe how synthesized populations are essential in real-scale social simulations (RSSS), and the current situation of the population synthesis for whole populations in Japan. RSSS is simulations using the real number of populations or households in social simulations. This paper describes how we have completed to synthesize multiple sets of populations based on the statistics of each local government in Japanese national census in 2000,2005,2010 and 2015. We have started to distribute those multiple sets of the synthesized populations for researchers of RSSSs in Japan. In distributing the synthesized populations, we should set some regulations in order to protect personal or private information in the synthesized populations.
{"title":"Distribution of Synthetic Populations of Japan for Social Scientists and Social Simulation Researchers","authors":"T. Murata, Takuya Harada, Manabu Ichika Wa, Yusuke Goto, Lee Hao, S. Date, M. Munetomo, Akiyoshi Sugiki","doi":"10.1109/ICMLC48188.2019.8949245","DOIUrl":"https://doi.org/10.1109/ICMLC48188.2019.8949245","url":null,"abstract":"In this paper, we describe how synthesized populations are essential in real-scale social simulations (RSSS), and the current situation of the population synthesis for whole populations in Japan. RSSS is simulations using the real number of populations or households in social simulations. This paper describes how we have completed to synthesize multiple sets of populations based on the statistics of each local government in Japanese national census in 2000,2005,2010 and 2015. We have started to distribute those multiple sets of the synthesized populations for researchers of RSSSs in Japan. In distributing the synthesized populations, we should set some regulations in order to protect personal or private information in the synthesized populations.","PeriodicalId":221349,"journal":{"name":"2019 International Conference on Machine Learning and Cybernetics (ICMLC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125829625","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 : 2019-07-01DOI: 10.1109/ICMLC48188.2019.8949290
Jianjun Zhang, Ting Wang, Wing W. Y. Ng, Shuai Zhang, C. Nugent
Undersampling the dataset to rebalance the class distribution is effective to handle class imbalance problems. However, randomly removing majority examples via a uniform distribution may lead to unnecessary information loss. This would result in performance deterioration of classifiers trained using this rebalanced dataset. On the other hand, examples have different sensitivities with respect to class imbalance. Higher sensitivity means that this example is more easily to be affected by class imbalance, which can be used to guide the selection of examples to rebalance the class distribution and to boost the classifier performance. Therefore, in this paper, we propose a novel undersampling method, the UnderSampling using Sensitivity (USS), based on sensitivity of each majority example. Examples with low sensitivities are noisy or safe examples while examples with high sensitivities are borderline examples. In USS, majority examples with higher sensitivities are more likely to be selected. Experiments on 20 datasets confirm the superiority of the USS against one baseline method and five resampling methods.
{"title":"Undersampling Near Decision Boundary for Imbalance Problems","authors":"Jianjun Zhang, Ting Wang, Wing W. Y. Ng, Shuai Zhang, C. Nugent","doi":"10.1109/ICMLC48188.2019.8949290","DOIUrl":"https://doi.org/10.1109/ICMLC48188.2019.8949290","url":null,"abstract":"Undersampling the dataset to rebalance the class distribution is effective to handle class imbalance problems. However, randomly removing majority examples via a uniform distribution may lead to unnecessary information loss. This would result in performance deterioration of classifiers trained using this rebalanced dataset. On the other hand, examples have different sensitivities with respect to class imbalance. Higher sensitivity means that this example is more easily to be affected by class imbalance, which can be used to guide the selection of examples to rebalance the class distribution and to boost the classifier performance. Therefore, in this paper, we propose a novel undersampling method, the UnderSampling using Sensitivity (USS), based on sensitivity of each majority example. Examples with low sensitivities are noisy or safe examples while examples with high sensitivities are borderline examples. In USS, majority examples with higher sensitivities are more likely to be selected. Experiments on 20 datasets confirm the superiority of the USS against one baseline method and five resampling methods.","PeriodicalId":221349,"journal":{"name":"2019 International Conference on Machine Learning and Cybernetics (ICMLC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130559211","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 : 2019-07-01DOI: 10.1109/ICMLC48188.2019.8949318
Yan Li, Jing Zhang, Qiang He, Siyuan Liu, Lujing Huo
In rough set theory, two crisp sets (i.e., the lower and upper approximates of a target concept) is used to describe uncertainties in given information systems. However, the traditional rough set models are built based on equivalence relations which do not consider the preference relationship of attribute values. Dominance relation-based rough set approach effectively solve this problem which uses dominance relations to substitute equivalence relations to deal with ordered data. In this kind of approach, the computing of dominance class is a necessary step to attribute reduction which is very time-consuming. In order to reduce the computational cost in calculating dominance classes, this paper presents a method to compute dominance classes by gradually reducing the search space in the domain. The corresponding algorithm is proposed. In each step of the algorithm, the inferior classes of the objects in a given information system are removed in the universe with the increase of the attributes. Experiments using six UCI data show that the proposed method improves the efficiency of computing dominance classes with the increasing of attributes and objects.
{"title":"An Acceleration Method for Computing Dominace Classes in Ordered Information System","authors":"Yan Li, Jing Zhang, Qiang He, Siyuan Liu, Lujing Huo","doi":"10.1109/ICMLC48188.2019.8949318","DOIUrl":"https://doi.org/10.1109/ICMLC48188.2019.8949318","url":null,"abstract":"In rough set theory, two crisp sets (i.e., the lower and upper approximates of a target concept) is used to describe uncertainties in given information systems. However, the traditional rough set models are built based on equivalence relations which do not consider the preference relationship of attribute values. Dominance relation-based rough set approach effectively solve this problem which uses dominance relations to substitute equivalence relations to deal with ordered data. In this kind of approach, the computing of dominance class is a necessary step to attribute reduction which is very time-consuming. In order to reduce the computational cost in calculating dominance classes, this paper presents a method to compute dominance classes by gradually reducing the search space in the domain. The corresponding algorithm is proposed. In each step of the algorithm, the inferior classes of the objects in a given information system are removed in the universe with the increase of the attributes. Experiments using six UCI data show that the proposed method improves the efficiency of computing dominance classes with the increasing of attributes and objects.","PeriodicalId":221349,"journal":{"name":"2019 International Conference on Machine Learning and Cybernetics (ICMLC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126793686","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 : 2019-07-01DOI: 10.1109/ICMLC48188.2019.8949324
Yun Zhang, Wenxiang Chen, Han Liu, Jinhua Liu, Hui Du
Binocular stereo matching aims to obtain disparities from two very close views. Existing stereo matching methods may cause false matching when there are much image noise and disparity discontinuities. This paper proposes a novel binocular stereo matching algorithm based on SAD and improved Census transformation. We first perform improved Census transformation, and then we get the matching costs by combining SAD and improved Census transformation. Finally we cluster the matching costs and calculate the disparities. To generate better disparities, we further propose the improved bilateral and selective filters to enhance the accuracy of disparities. Experimental results show that our binocular stereo matching can produce more accurate and complete disparities, and it works well in complex scenes with irregular shapes and more objects, thus it has wide applications in stereoscopic image processing.
{"title":"Efficient Binocular Stereo Matching Based on Sad and Improved Census Transformation","authors":"Yun Zhang, Wenxiang Chen, Han Liu, Jinhua Liu, Hui Du","doi":"10.1109/ICMLC48188.2019.8949324","DOIUrl":"https://doi.org/10.1109/ICMLC48188.2019.8949324","url":null,"abstract":"Binocular stereo matching aims to obtain disparities from two very close views. Existing stereo matching methods may cause false matching when there are much image noise and disparity discontinuities. This paper proposes a novel binocular stereo matching algorithm based on SAD and improved Census transformation. We first perform improved Census transformation, and then we get the matching costs by combining SAD and improved Census transformation. Finally we cluster the matching costs and calculate the disparities. To generate better disparities, we further propose the improved bilateral and selective filters to enhance the accuracy of disparities. Experimental results show that our binocular stereo matching can produce more accurate and complete disparities, and it works well in complex scenes with irregular shapes and more objects, thus it has wide applications in stereoscopic image processing.","PeriodicalId":221349,"journal":{"name":"2019 International Conference on Machine Learning and Cybernetics (ICMLC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127388135","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：