Aiming at the complex dependence of water quality data in space and time, we propose a GCN-Seq2Seq model for surface water quality prediction. The model uses Graph Convolutional Network (GCN) to capture the spatial feature of water quality monitoring sites, uses the sequence to sequence (Seq2Seq) model constructed by GRU to extract the temporal feature of the water quality data sequence, and predicts multi-step water quality time series. Experiments were carried out with data from 6 water quality monitoring stations in the Huangshui River and surrounding areas in Xining City, Qinghai Province from November 2020 to June 2021, and compared with the baseline model. experimental results show that the proposed model can effectively improve the accuracy of multi-step prediction of surface water quality.
针对地表水水质数据在空间和时间上的复杂依赖性，提出了一种用于地表水水质预测的GCN-Seq2Seq模型。该模型利用图卷积网络(Graph Convolutional Network, GCN)捕捉水质监测点的空间特征，利用GRU构建的序列到序列(sequence to sequence, Seq2Seq)模型提取水质数据序列的时间特征，并对多步水质时间序列进行预测。利用青海省西宁市湟水河及周边地区6个水质监测站2020年11月至2021年6月的数据进行实验，并与基线模型进行对比。实验结果表明，该模型能有效提高地表水水质多步预测的精度。
{"title":"GCN-Seq2Seq: A Spatio-Temporal feature-fused model for surface water quality prediction","authors":"Ying Chen, Ping Yang, Chengxu Ye, Zhikun Miao","doi":"10.1145/3507548.3507597","DOIUrl":"https://doi.org/10.1145/3507548.3507597","url":null,"abstract":"Aiming at the complex dependence of water quality data in space and time, we propose a GCN-Seq2Seq model for surface water quality prediction. The model uses Graph Convolutional Network (GCN) to capture the spatial feature of water quality monitoring sites, uses the sequence to sequence (Seq2Seq) model constructed by GRU to extract the temporal feature of the water quality data sequence, and predicts multi-step water quality time series. Experiments were carried out with data from 6 water quality monitoring stations in the Huangshui River and surrounding areas in Xining City, Qinghai Province from November 2020 to June 2021, and compared with the baseline model. experimental results show that the proposed model can effectively improve the accuracy of multi-step prediction of surface water quality.","PeriodicalId":414908,"journal":{"name":"Proceedings of the 2021 5th International Conference on Computer Science and Artificial Intelligence","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130873026","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}
To address the lack of up-to-date smoke detection datasets, we have compiled and labeled a variety smoke detection dataset called SM-dataset. This dataset contains a total of 11596 smoke images from natural scenes. Meanwhile, we introduce a new version of YOLO with better performance, which we call SM-YOLO. SM-YOLO builds on the original model of YOLOv5m, reduces the original three outputs to two, streamlines the original network structure and improves the loss of the original network. Compared with YOLOv5m, SM-YOLO has only 75% of the trainable parameters, but improves mAP@.5 by relative 2%, and reduces the inference time of a single frame from 7.3 ms to 6.6 ms, which effectively improves the speed of smoke detection.
{"title":"SM-YOLO: A Model for Real-Time Smoke Detection","authors":"Zhen Yang, Han Yu, Lei Xu, Fan Yang, Zhijian Yin","doi":"10.1145/3507548.3507554","DOIUrl":"https://doi.org/10.1145/3507548.3507554","url":null,"abstract":"To address the lack of up-to-date smoke detection datasets, we have compiled and labeled a variety smoke detection dataset called SM-dataset. This dataset contains a total of 11596 smoke images from natural scenes. Meanwhile, we introduce a new version of YOLO with better performance, which we call SM-YOLO. SM-YOLO builds on the original model of YOLOv5m, reduces the original three outputs to two, streamlines the original network structure and improves the loss of the original network. Compared with YOLOv5m, SM-YOLO has only 75% of the trainable parameters, but improves mAP@.5 by relative 2%, and reduces the inference time of a single frame from 7.3 ms to 6.6 ms, which effectively improves the speed of smoke detection.","PeriodicalId":414908,"journal":{"name":"Proceedings of the 2021 5th International Conference on Computer Science and Artificial Intelligence","volume":"415 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133964381","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}
Video question answering (VideoQA) is an intriguing topic in the field of visual language. Most of the current VideoQA models directly harness the global video information to answer questions. However, in VideoQA task, the answers associated with the questions merely appear in a few video contents, and other contents are invalid and redundant information. Therefore, VideoQA is vulnerable to be interfered by a large number of irrelevant contents. To address this challenge, we propose a Keyword-aware Multi-modal Enhancement Attention model for VideoQA. Specifically, a multi-factor keyword extraction (MFKE) algorithm is proposed to emphasize the crucial information in multimodal feature extraction. Furthermore, based on attention mechanisms, a keyword-aware enhancement attention (KAEA) module is designed to correlate the information associated with multiple modalities and fuse multimodal features. The experimental results on publicly available large VideoQA datasets, namely TVQA+ and LifeQA, demonstrate the effectiveness of our model.
{"title":"Keyword-aware Multi-modal Enhancement Attention for Video Question Answering","authors":"Duo Chen, Fuwei Zhang, Shirou Ou, Ruomei Wang","doi":"10.1145/3507548.3507567","DOIUrl":"https://doi.org/10.1145/3507548.3507567","url":null,"abstract":"Video question answering (VideoQA) is an intriguing topic in the field of visual language. Most of the current VideoQA models directly harness the global video information to answer questions. However, in VideoQA task, the answers associated with the questions merely appear in a few video contents, and other contents are invalid and redundant information. Therefore, VideoQA is vulnerable to be interfered by a large number of irrelevant contents. To address this challenge, we propose a Keyword-aware Multi-modal Enhancement Attention model for VideoQA. Specifically, a multi-factor keyword extraction (MFKE) algorithm is proposed to emphasize the crucial information in multimodal feature extraction. Furthermore, based on attention mechanisms, a keyword-aware enhancement attention (KAEA) module is designed to correlate the information associated with multiple modalities and fuse multimodal features. The experimental results on publicly available large VideoQA datasets, namely TVQA+ and LifeQA, demonstrate the effectiveness of our model.","PeriodicalId":414908,"journal":{"name":"Proceedings of the 2021 5th International Conference on Computer Science and Artificial Intelligence","volume":"74 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133636718","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 explosion in the popularity of smartphones over the previous decade, mobile malware appears to be unavoidable. Because Android is an open platform that is fast dominating other rival platforms (e.g. iOS) in the mobile smart device industry, Android malware has been much more widespread. Recent Android malware developers have more advanced capabilities when building their malicious apps, which make the apps themselves much more difficult to detect using conventional methods. In our paper, we proposed a hybrid machine learning classification and clustering algorithm to detect recent Android malware. The proposed algorithm performs better than the state-of-art algorithms with both F1-score and recall of 0.9944. More importantly, the top features returned by our algorithm clearly explain the important factors in the detection task. They can not only be used for enhanced Android malware detection but also quicker white-box analysis by means of more interpretable results.
{"title":"Hybrid Classification and Clustering Algorithm on Recent Android Malware Detection","authors":"jiezhong xiao, Qian Han, Yumeng Gao","doi":"10.1145/3507548.3507586","DOIUrl":"https://doi.org/10.1145/3507548.3507586","url":null,"abstract":"With the explosion in the popularity of smartphones over the previous decade, mobile malware appears to be unavoidable. Because Android is an open platform that is fast dominating other rival platforms (e.g. iOS) in the mobile smart device industry, Android malware has been much more widespread. Recent Android malware developers have more advanced capabilities when building their malicious apps, which make the apps themselves much more difficult to detect using conventional methods. In our paper, we proposed a hybrid machine learning classification and clustering algorithm to detect recent Android malware. The proposed algorithm performs better than the state-of-art algorithms with both F1-score and recall of 0.9944. More importantly, the top features returned by our algorithm clearly explain the important factors in the detection task. They can not only be used for enhanced Android malware detection but also quicker white-box analysis by means of more interpretable results.","PeriodicalId":414908,"journal":{"name":"Proceedings of the 2021 5th International Conference on Computer Science and Artificial Intelligence","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133499578","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 shape of the Chinese character should be square and not slanted in its overall presentation. Beginners of Chinese as a foreign language (CFL) often tend to write slanted characters as they have not yet fully grasped the writing techniques of the strokes and the relationship between them. Slant deviation in handwritten characters is usually assessed manually, which is time-consuming and subjective as there are no quantitative criteria. Existing methods of computing the slant membership of Chinese characters are mainly based on the angle of individual strokes, ignoring other conformational factors that affect the overall slant of the character. This paper proposes a slant membership computation approach for handwritten Chinese characters based on elliptical enclosing shapes, with the aim of computing the slant membership that reflects the combination of all Chinese strokes. A knowledge base is also constructed to label the slant information of standard template characters, and the slant membership of handwritten characters is computed by comparing the differences between them with the template characters in the knowledge base. Experiments conducted with digital ink character data from CFL beginners proved that the proposed approach is effective.
{"title":"Computing the Slant Degree of Digital Ink Chinese Characters Handwritten by CFL Beginners Based on Elliptical Enclosing Shape","authors":"Yun Lai, Xiwen Zhang","doi":"10.1145/3507548.3507562","DOIUrl":"https://doi.org/10.1145/3507548.3507562","url":null,"abstract":"∗The shape of the Chinese character should be square and not slanted in its overall presentation. Beginners of Chinese as a foreign language (CFL) often tend to write slanted characters as they have not yet fully grasped the writing techniques of the strokes and the relationship between them. Slant deviation in handwritten characters is usually assessed manually, which is time-consuming and subjective as there are no quantitative criteria. Existing methods of computing the slant membership of Chinese characters are mainly based on the angle of individual strokes, ignoring other conformational factors that affect the overall slant of the character. This paper proposes a slant membership computation approach for handwritten Chinese characters based on elliptical enclosing shapes, with the aim of computing the slant membership that reflects the combination of all Chinese strokes. A knowledge base is also constructed to label the slant information of standard template characters, and the slant membership of handwritten characters is computed by comparing the differences between them with the template characters in the knowledge base. Experiments conducted with digital ink character data from CFL beginners proved that the proposed approach is effective.","PeriodicalId":414908,"journal":{"name":"Proceedings of the 2021 5th International Conference on Computer Science and Artificial Intelligence","volume":"144 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122497344","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}
ECO based trackers have achieved excellent performance on visual object tracking. However, Illumination variation and other factors still are challenging research problems in the process of tracking. Moreover, traditional neural networks also face information loss during the transmission process. In this paper, we introduce a new feature fusion (HE, FHOG-Encoder) and update strategy of learning rate. We propose an encoder network to extract features, which consists of two convolutional layers and three residual units. In addition, we design an updating strategy of learning rate, by computing absolute difference of inter-frame pixel, to effectively update sample space model. Experiments on challenging benchmarks OTB-100 are carried out. Experimental results show that our tracker achieves superior performance in some special cases, compared with the original ECO tracker.
{"title":"Target tracking by improved ECO","authors":"JiaoJiao Xing, Xianmei Wang, Peng Hou","doi":"10.1145/3507548.3507555","DOIUrl":"https://doi.org/10.1145/3507548.3507555","url":null,"abstract":"ECO based trackers have achieved excellent performance on visual object tracking. However, Illumination variation and other factors still are challenging research problems in the process of tracking. Moreover, traditional neural networks also face information loss during the transmission process. In this paper, we introduce a new feature fusion (HE, FHOG-Encoder) and update strategy of learning rate. We propose an encoder network to extract features, which consists of two convolutional layers and three residual units. In addition, we design an updating strategy of learning rate, by computing absolute difference of inter-frame pixel, to effectively update sample space model. Experiments on challenging benchmarks OTB-100 are carried out. Experimental results show that our tracker achieves superior performance in some special cases, compared with the original ECO tracker.","PeriodicalId":414908,"journal":{"name":"Proceedings of the 2021 5th International Conference on Computer Science and Artificial Intelligence","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127817745","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}
Most density-based clustering algorithms are only suitable for spherical data set. When processing streamlined data sets without cluster centers, the clustering results have certain defects. In order to deal with the clustering problem of streamlined data sets, the concept of natural neighbors and outlier detection are combined, and a boundary-removing natural neighbor clustering (NNC_wbo) algorithm is proposed. First, establish the natural neighbor relationship between the KD tree search data, calculate the intra-group density and intra-group outlier degree of the data points, set the parameters to remove the boundary data; then use the natural neighbor relationship to obtain the preliminary clustering results; if after the preliminary clustering, There are small clusters composed of very few data points, and outliers are excluded.
{"title":"Natural Neighbor Clustering Algorithm without Boundary","authors":"Luzou Zhang, Yunjie Zhang, Yulin Wang","doi":"10.1145/3507548.3507584","DOIUrl":"https://doi.org/10.1145/3507548.3507584","url":null,"abstract":"Most density-based clustering algorithms are only suitable for spherical data set. When processing streamlined data sets without cluster centers, the clustering results have certain defects. In order to deal with the clustering problem of streamlined data sets, the concept of natural neighbors and outlier detection are combined, and a boundary-removing natural neighbor clustering (NNC_wbo) algorithm is proposed. First, establish the natural neighbor relationship between the KD tree search data, calculate the intra-group density and intra-group outlier degree of the data points, set the parameters to remove the boundary data; then use the natural neighbor relationship to obtain the preliminary clustering results; if after the preliminary clustering, There are small clusters composed of very few data points, and outliers are excluded.","PeriodicalId":414908,"journal":{"name":"Proceedings of the 2021 5th International Conference on Computer Science and Artificial Intelligence","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129295609","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}
{"title":"Proceedings of the 2021 5th International Conference on Computer Science and Artificial Intelligence","authors":"","doi":"10.1145/3507548","DOIUrl":"https://doi.org/10.1145/3507548","url":null,"abstract":"","PeriodicalId":414908,"journal":{"name":"Proceedings of the 2021 5th International Conference on Computer Science and Artificial Intelligence","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1900-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128318032","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：