Blockchain is a decentralized technology proposed by Satoshi Nakamoto in 2008, without relying on trust, irrevocable and modified, based on a consensus mechanism. Blockchain technology was not created out of thin air, but was born as the underlying technology of bitcoin, a digital currency. Now it has received widespread attention for its implementation in distributed ledger. However, in the current era, the application of blockchain technology in actual scenarios has not yet matured, mainly due to some characteristics of blockchain technology itself. After weighing its advantages and disadvantages, it is increasingly becoming its limitation. One of the main aspects is the poor scalability of the blockchain, which not only limits the functional expansion of the blockchain and makes it impossible to apply the blockchain technology to a wider range of scenarios, but also limits the throughput of the blockchain system to a certain extent Promotion. Taking into account the characteristics and bottlenecks of the above-mentioned blockchain technology, different from previous studies, this article from the perspective of blockchain expansion, introduces some existing expansion technologies in solving the transaction rate of the blockchain, and analyzes its principles and feasibility. Through the classification and comparison of different expansion technologies, the advantages and possible disadvantages of various expansion technologies are described. Finally, on the basis of fully understanding the existing blockchain expansion technology, the outlook for the future development of blockchain expansion technology is proposed, with a view to providing some suggestions for the development of the blockchain expansion technology.
{"title":"The Challenge and Prospect of Scalability of Blockchain Technology","authors":"Lizhi Wang","doi":"10.1145/3507548.3507594","DOIUrl":"https://doi.org/10.1145/3507548.3507594","url":null,"abstract":"Blockchain is a decentralized technology proposed by Satoshi Nakamoto in 2008, without relying on trust, irrevocable and modified, based on a consensus mechanism. Blockchain technology was not created out of thin air, but was born as the underlying technology of bitcoin, a digital currency. Now it has received widespread attention for its implementation in distributed ledger. However, in the current era, the application of blockchain technology in actual scenarios has not yet matured, mainly due to some characteristics of blockchain technology itself. After weighing its advantages and disadvantages, it is increasingly becoming its limitation. One of the main aspects is the poor scalability of the blockchain, which not only limits the functional expansion of the blockchain and makes it impossible to apply the blockchain technology to a wider range of scenarios, but also limits the throughput of the blockchain system to a certain extent Promotion. Taking into account the characteristics and bottlenecks of the above-mentioned blockchain technology, different from previous studies, this article from the perspective of blockchain expansion, introduces some existing expansion technologies in solving the transaction rate of the blockchain, and analyzes its principles and feasibility. Through the classification and comparison of different expansion technologies, the advantages and possible disadvantages of various expansion technologies are described. Finally, on the basis of fully understanding the existing blockchain expansion technology, the outlook for the future development of blockchain expansion technology is proposed, with a view to providing some suggestions for the development of the blockchain expansion technology.","PeriodicalId":414908,"journal":{"name":"Proceedings of the 2021 5th International Conference on Computer Science and Artificial Intelligence","volume":"34 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":"123420930","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}
Zhuxin Xue, Yang Bai, Haixin Wang, Chenyu He, Jian Tan
Human-computer interaction is a technology to study the relationship between users and systems. Good user experience can greatly enhance user stickiness. With the development of Internet technology, users begin to face a variety of systems, but the interaction methods involved are different. Learning new interaction means an increase in user learning costs. As a part of product design, the design of interaction also needs to cover special user groups as much as possible, such as blind people. Thus, it is necessary to study whether different ways of interaction can migrate to each other. Traditional research on user experience mainly focuses on qualitative aspects, such as questionnaire survey. In this paper, we propose a quantitative model to evaluate the portability of interaction. For software interaction design, the macro concept of user experience is quantified by different dimensions, and a unified index model and calculation method are output to guide and evaluate the portability of different software interactions.
{"title":"Research on human-computer interaction portability evaluation model in complex environment","authors":"Zhuxin Xue, Yang Bai, Haixin Wang, Chenyu He, Jian Tan","doi":"10.1145/3507548.3507615","DOIUrl":"https://doi.org/10.1145/3507548.3507615","url":null,"abstract":"Human-computer interaction is a technology to study the relationship between users and systems. Good user experience can greatly enhance user stickiness. With the development of Internet technology, users begin to face a variety of systems, but the interaction methods involved are different. Learning new interaction means an increase in user learning costs. As a part of product design, the design of interaction also needs to cover special user groups as much as possible, such as blind people. Thus, it is necessary to study whether different ways of interaction can migrate to each other. Traditional research on user experience mainly focuses on qualitative aspects, such as questionnaire survey. In this paper, we propose a quantitative model to evaluate the portability of interaction. For software interaction design, the macro concept of user experience is quantified by different dimensions, and a unified index model and calculation method are output to guide and evaluate the portability of different software interactions.","PeriodicalId":414908,"journal":{"name":"Proceedings of the 2021 5th International Conference on Computer Science and Artificial Intelligence","volume":"211 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":"123018770","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}
Aiming at determining the weights of selected investment targets to obtain higher and more stable investment returns, a Temporal Convolution Networks (TCN) based portfolio model, namely, TCNportfolio model is proposed. TCN-portfolio model combines TCNbased time series processing and MLP-based cross-sectional data processing, and finally outputs the investment target weights which changes every ten trading days. We optimize the TCN-portfolio model using a Multi-Objective Genetic Algorithm (MOGA) which optimizes the rate of return and variance at the same time. The component stocks of Shanghai Securities Composite 50 index (SSEC 50) are selected as the investment targets. Experimental results on the test sets reveal that TCN-portfolio model performs well. Its average daily return rate is obviously greater than those of SSEC and SSEC 50, and the cumulative return rate of TCN-portfolio model is always greater than those of SSEC and SSEC 50 on the test data set.
{"title":"The Portfolio Model Based on Temporal Convolution Networks and the Empirical Research on Chinese Stock Market","authors":"Rui Zhang, Zuoquan Zhang, Marui Du, Xiaomin Wang","doi":"10.1145/3507548.3507593","DOIUrl":"https://doi.org/10.1145/3507548.3507593","url":null,"abstract":"Aiming at determining the weights of selected investment targets to obtain higher and more stable investment returns, a Temporal Convolution Networks (TCN) based portfolio model, namely, TCNportfolio model is proposed. TCN-portfolio model combines TCNbased time series processing and MLP-based cross-sectional data processing, and finally outputs the investment target weights which changes every ten trading days. We optimize the TCN-portfolio model using a Multi-Objective Genetic Algorithm (MOGA) which optimizes the rate of return and variance at the same time. The component stocks of Shanghai Securities Composite 50 index (SSEC 50) are selected as the investment targets. Experimental results on the test sets reveal that TCN-portfolio model performs well. Its average daily return rate is obviously greater than those of SSEC and SSEC 50, and the cumulative return rate of TCN-portfolio model is always greater than those of SSEC and SSEC 50 on the test data set.","PeriodicalId":414908,"journal":{"name":"Proceedings of the 2021 5th International Conference on Computer Science and Artificial Intelligence","volume":"7 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":"126879775","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}
Bundit Manaskasemsak, Sarita Puttitanun, Jirateep Tantisuwankul, A. Rungsawang
The rise of user-generated content on the Internet today has led to the problem of data overload. Therefore, recommender systems have been introduced in various social platforms to automatically serve interesting content to users. Pantip.com is the most popular Thai Internet forum where people can discuss ideas, tips, and news on a variety of topics. Although Pantip has many recommendation services, these are not specific for individual users. In this paper, we proposed a personalized thread recommender system that is applicable to the Pantip site. The approach finds out appropriate threads for each user based on three aspects: user interests, thread trends, and thread freshness along with the analysis in changing of user behavior over time. We conducted experiments on the Pantip clickstream dataset and evaluated the performance by real users. Experimental results show that the proposed approach recommends threads that are significantly more satisfying for users than the baseline approaches.
{"title":"Personalized Thread Recommendation on Thai Internet Forum","authors":"Bundit Manaskasemsak, Sarita Puttitanun, Jirateep Tantisuwankul, A. Rungsawang","doi":"10.1145/3507548.3507589","DOIUrl":"https://doi.org/10.1145/3507548.3507589","url":null,"abstract":"The rise of user-generated content on the Internet today has led to the problem of data overload. Therefore, recommender systems have been introduced in various social platforms to automatically serve interesting content to users. Pantip.com is the most popular Thai Internet forum where people can discuss ideas, tips, and news on a variety of topics. Although Pantip has many recommendation services, these are not specific for individual users. In this paper, we proposed a personalized thread recommender system that is applicable to the Pantip site. The approach finds out appropriate threads for each user based on three aspects: user interests, thread trends, and thread freshness along with the analysis in changing of user behavior over time. We conducted experiments on the Pantip clickstream dataset and evaluated the performance by real users. Experimental results show that the proposed approach recommends threads that are significantly more satisfying for users than the baseline approaches.","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":"133608085","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}
Music visualization is a visual art form for understanding, analyzing and comparing the internal structure and expressive features of music. It meets the aesthetic demand of the masses in the digital age. This paper reviews the development and research status of the music visualization literature in the past 20 years, comprehensively analyzes the research process and current hotspots of music visualization, and speculates the future development trend. We have used Web of Science (WoS) and China National Knowledge Infrastructure (CNKI) as data sources, used CiteSpace software to compare and analyze the year, country, subject distribution and hot keywords of music visualization literature at China and the world from 2000 to 2020 by the method of Mapping Knowledge Domain. The results show that the research on music visualization at China and other countries is showing an upward trend, and it presents the characteristics of multi-disciplinary integration. Different application scenarios, research methods and development stages lead to different research hotspots between different countries. The shortcomings of Chinese research in this field lies in that the research content needs to be deepened, the interdisciplinary content needs to be integrated, and applications of music visualization needs to be popularized.
音乐可视化是一种理解、分析和比较音乐的内在结构和表现特征的视觉艺术形式。它满足了数字时代大众的审美需求。本文回顾了近20年来音乐可视化文献的发展和研究现状，综合分析了音乐可视化的研究过程和当前热点，并对未来的发展趋势进行了推测。以WoS (Web of Science)和CNKI (China National Knowledge Infrastructure)为数据源，利用CiteSpace软件，采用知识图谱的方法，对2000 - 2020年中国和世界音乐可视化文献的年份、国家、学科分布和热点关键词进行了比较分析。结果表明，国内外对音乐可视化的研究呈现出上升趋势，呈现出多学科融合的特点。不同的应用场景、研究方法和发展阶段导致不同国家的研究热点不同。中国在该领域研究的不足在于研究内容有待深化，跨学科内容有待整合，音乐可视化应用有待推广。
{"title":"Comparative Study of Music Visualization based on CiteSpace at China and the World","authors":"Hai-Yan Zheng, Zhengqing Jiang","doi":"10.1145/3507548.3507604","DOIUrl":"https://doi.org/10.1145/3507548.3507604","url":null,"abstract":"Music visualization is a visual art form for understanding, analyzing and comparing the internal structure and expressive features of music. It meets the aesthetic demand of the masses in the digital age. This paper reviews the development and research status of the music visualization literature in the past 20 years, comprehensively analyzes the research process and current hotspots of music visualization, and speculates the future development trend. We have used Web of Science (WoS) and China National Knowledge Infrastructure (CNKI) as data sources, used CiteSpace software to compare and analyze the year, country, subject distribution and hot keywords of music visualization literature at China and the world from 2000 to 2020 by the method of Mapping Knowledge Domain. The results show that the research on music visualization at China and other countries is showing an upward trend, and it presents the characteristics of multi-disciplinary integration. Different application scenarios, research methods and development stages lead to different research hotspots between different countries. The shortcomings of Chinese research in this field lies in that the research content needs to be deepened, the interdisciplinary content needs to be integrated, and applications of music visualization needs to be popularized.","PeriodicalId":414908,"journal":{"name":"Proceedings of the 2021 5th International Conference on Computer Science and Artificial Intelligence","volume":"21 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":"115460282","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}
Dongxue Bao, Donghong Qin, Xianye Liang, Lila Hong
Abstract: Aiming at short texts lacking contextual information, large amount of text data, sparse features, and traditional text feature representations that cannot dynamically obtain the key classification information of a word polysemous and contextual semantics. this paper proposes a pre-trained language model based on BERT. The network model B-BAtt-MPC (BERT-BiLSTM-Attention-Max-Pooling-Concat) that integrates BiLSTM, Attention mechanism and Max-Pooling mechanism. Firstly, obtain multi-dimensional and rich feature information such as text context semantics, grammar, and context through the BERT model; Secondly, use the BERT output vector to obtain the most important feature information worth noting through the BiLSTM, Attention layer and Max-Pooling layer; In order to optimize the classification model, the BERT and BiLSTM output vectors are fused and input into Max-Pooling; Finally, the classification results are obtained by fusing two feature vectors with Max-Pooling. The experimental results of two data sets show that the model proposed in this paper can obtain the importance and key rich semantic features of short text classification, and can improve the text classification effect.
摘要:针对缺乏上下文信息的短文本、文本数据量大、特征稀疏、传统文本特征表示不能动态获取词的多义和上下文语义的关键分类信息等问题。本文提出了一种基于BERT的预训练语言模型。集成了BiLSTM、Attention机制和Max-Pooling机制的网络模型b - bat - mpc (BERT-BiLSTM-Attention-Max-Pooling-Concat)。首先，通过BERT模型获得文本上下文语义、语法、上下文等多维、丰富的特征信息;其次，利用BERT输出向量，通过BiLSTM、Attention层和Max-Pooling层获得最重要的值得注意的特征信息;为了优化分类模型，将BERT和BiLSTM输出向量融合并输入到Max-Pooling中;最后，利用Max-Pooling对两个特征向量进行融合，得到分类结果。两个数据集的实验结果表明，本文提出的模型能够获得短文本分类的重要性和关键丰富的语义特征，能够提高文本分类效果。
{"title":"Short Text Classification Model Based on BERT and Fusion Network","authors":"Dongxue Bao, Donghong Qin, Xianye Liang, Lila Hong","doi":"10.1145/3507548.3507574","DOIUrl":"https://doi.org/10.1145/3507548.3507574","url":null,"abstract":"Abstract: Aiming at short texts lacking contextual information, large amount of text data, sparse features, and traditional text feature representations that cannot dynamically obtain the key classification information of a word polysemous and contextual semantics. this paper proposes a pre-trained language model based on BERT. The network model B-BAtt-MPC (BERT-BiLSTM-Attention-Max-Pooling-Concat) that integrates BiLSTM, Attention mechanism and Max-Pooling mechanism. Firstly, obtain multi-dimensional and rich feature information such as text context semantics, grammar, and context through the BERT model; Secondly, use the BERT output vector to obtain the most important feature information worth noting through the BiLSTM, Attention layer and Max-Pooling layer; In order to optimize the classification model, the BERT and BiLSTM output vectors are fused and input into Max-Pooling; Finally, the classification results are obtained by fusing two feature vectors with Max-Pooling. The experimental results of two data sets show that the model proposed in this paper can obtain the importance and key rich semantic features of short text classification, and can improve the text classification effect.","PeriodicalId":414908,"journal":{"name":"Proceedings of the 2021 5th International Conference on Computer Science and Artificial Intelligence","volume":"22 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":"125693208","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}
Bin-Bin Yang, Shengjie Zhao, Kenan Ye, Rongqing Zhang
Ordinal regression is a typical deep learning problem, which involves inherently ordered labels that are common in practical applications, especially in medical diagnosis tasks. To overcome the neglect of ordered or non-stationary property by merely exploiting classification or regression, quadratic weighted kappa (QWK) is proposed to be employed in the QWK loss function design as an efficient evaluation metric for ordinal regression. However, the paradox that kappa will be higher with an asymmetrical marginal histogram leads the QWK loss function to get the local optimal solution with all-zero-column in the confusion matrices during training. In practice, the all-zero column problem will result in a certain category not being detected at all, which can have serious consequences for the exclusion of pathology. To address this limitation, a new form of penalty term is proposed for the QWK loss function by penalizing the distance of marginal histogram to effectively avoid all-zero-column of the models. The experiments on the category-imbalanced datasets demonstrate that our penalty terms solve all-zero-column problem. On Adience dataset our penalty terms achieve 0.915 QWK, 0.446 MAE and 0.612 accuracy, while on DR dataset our penalty terms achieve 0.744 QWK, 0.281 MAE and 0.810 accuracy. Besides, experiments on the category-balanced datasets HCI show that our penalty terms achieve 0.810 QWK, 0.499 MAE and 0.610 accuracy.
{"title":"Distribution Consistency Penalty in the Quadratic Kappa Loss for Ordinal Regression of Imbalanced Datasets","authors":"Bin-Bin Yang, Shengjie Zhao, Kenan Ye, Rongqing Zhang","doi":"10.1145/3507548.3507612","DOIUrl":"https://doi.org/10.1145/3507548.3507612","url":null,"abstract":"Ordinal regression is a typical deep learning problem, which involves inherently ordered labels that are common in practical applications, especially in medical diagnosis tasks. To overcome the neglect of ordered or non-stationary property by merely exploiting classification or regression, quadratic weighted kappa (QWK) is proposed to be employed in the QWK loss function design as an efficient evaluation metric for ordinal regression. However, the paradox that kappa will be higher with an asymmetrical marginal histogram leads the QWK loss function to get the local optimal solution with all-zero-column in the confusion matrices during training. In practice, the all-zero column problem will result in a certain category not being detected at all, which can have serious consequences for the exclusion of pathology. To address this limitation, a new form of penalty term is proposed for the QWK loss function by penalizing the distance of marginal histogram to effectively avoid all-zero-column of the models. The experiments on the category-imbalanced datasets demonstrate that our penalty terms solve all-zero-column problem. On Adience dataset our penalty terms achieve 0.915 QWK, 0.446 MAE and 0.612 accuracy, while on DR dataset our penalty terms achieve 0.744 QWK, 0.281 MAE and 0.810 accuracy. Besides, experiments on the category-balanced datasets HCI show that our penalty terms achieve 0.810 QWK, 0.499 MAE and 0.610 accuracy.","PeriodicalId":414908,"journal":{"name":"Proceedings of the 2021 5th International Conference on Computer Science and Artificial Intelligence","volume":"22 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":"132031111","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 world has faced a severe challenge since the breakout of the novel Coronavirus-2019 (COVID-19) has started for more than one year. With the mutation of the virus, the measures of epidemic prevention are keeping upgrading. Various vaccines have been created and brought into operation. To accurately describe and predict the spread of COVID-19, we improve the traditional Susceptible-Exposed-Infected-Removed-Dead model(SEIRD), forecast the development of COVID-19 based on small-world network. A small-world network is a type of mathematical graph in which most nodes are not neighbors of one another, but the neighbors of any given node are likely to be neighbors of each other, and most nodes can be reached from every other node by a small number of hops or steps. We introduce new parameters, Vaccination(V) and Quarantine(Q), into this model. Based on this, through regressing and analyzing the epidemic in the UK, we get the simulation that fits well with the observed data in other countries.
{"title":"Forecast of the Development of COVID-19 Based on the Small-World Network","authors":"Xingye Bu, Naijie Gu","doi":"10.1145/3507548.3507575","DOIUrl":"https://doi.org/10.1145/3507548.3507575","url":null,"abstract":"This world has faced a severe challenge since the breakout of the novel Coronavirus-2019 (COVID-19) has started for more than one year. With the mutation of the virus, the measures of epidemic prevention are keeping upgrading. Various vaccines have been created and brought into operation. To accurately describe and predict the spread of COVID-19, we improve the traditional Susceptible-Exposed-Infected-Removed-Dead model(SEIRD), forecast the development of COVID-19 based on small-world network. A small-world network is a type of mathematical graph in which most nodes are not neighbors of one another, but the neighbors of any given node are likely to be neighbors of each other, and most nodes can be reached from every other node by a small number of hops or steps. We introduce new parameters, Vaccination(V) and Quarantine(Q), into this model. Based on this, through regressing and analyzing the epidemic in the UK, we get the simulation that fits well with the observed data in other countries.","PeriodicalId":414908,"journal":{"name":"Proceedings of the 2021 5th International Conference on Computer Science and Artificial Intelligence","volume":"6 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":"130864257","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 prediction of program running time can be used to improve scheduling performance of distributed systems. In 2011, Google released a data set documenting the vast amount of information in the Google cluster. However, most of the existing running time prediction models only consider the coarse-grained characteristics of the running environment without considering the influence of the time series data of the running environment on the prediction results. Based on this, this paper innovatively proposes a model to predict the running time of the program, which predicts the future running time through historical information. At the same time, we also propose a new data processing and feature extraction scheme for Google cluster data sets. The results show that our model greatly outperforms the classical model on the Google cluster data set, and the root-mean-square error index of running time under different prediction modes is reduced by more than 60% and 40%, respectively. We hope that the model proposed in this paper can provide new research ideas for cloud computing system design.
{"title":"Use Machine Learning to Predict the Running Time of the Program","authors":"Xinyi Li, Yiyuan Wang, Ying Qian, Liang Dou","doi":"10.1145/3507548.3507588","DOIUrl":"https://doi.org/10.1145/3507548.3507588","url":null,"abstract":"The prediction of program running time can be used to improve scheduling performance of distributed systems. In 2011, Google released a data set documenting the vast amount of information in the Google cluster. However, most of the existing running time prediction models only consider the coarse-grained characteristics of the running environment without considering the influence of the time series data of the running environment on the prediction results. Based on this, this paper innovatively proposes a model to predict the running time of the program, which predicts the future running time through historical information. At the same time, we also propose a new data processing and feature extraction scheme for Google cluster data sets. The results show that our model greatly outperforms the classical model on the Google cluster data set, and the root-mean-square error index of running time under different prediction modes is reduced by more than 60% and 40%, respectively. We hope that the model proposed in this paper can provide new research ideas for cloud computing system design.","PeriodicalId":414908,"journal":{"name":"Proceedings of the 2021 5th International Conference on Computer Science and Artificial Intelligence","volume":"31 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":"121607015","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}
Christos G. Chadoulos, S. Moustakidis, D. Tsaopoulos, J. Theocharis
Multi-atlas based segmentation techniques have been proven to be effective in multiple automatic segmentation applications. However, mostly they rely on a non-deformable registration model followed by a voxel-wise classification process that incurs a large computational cost in terms of memory requirements and execution time. In this paper, a novel two-stage multi-atlas method is proposed, which combines constructively several concepts, including Semi-Supervised Learning (SSL), sparse graph constructions, voxel’s linear reconstructions via graph weights, and suitable sampling schemes for collecting data from target image and the atlas library. Representative global data sampled from target image are first classified according to SSL, using a newly proposed label propagation scheme. Next, out-of-sample data of yet unlabeled target voxels are iteratively generated through an iterative sampling based on mesh tetrahedralization. A thorough experimental investigation is conducted on 45 subjects provided by the publicly accessible Osteoarthritis Initiative (OAI) repository. Comparative analysis demonstrates that the proposed approach outperforms the existing state-of-the-art patch-based methods, across all evaluation metrics, exhibiting enhanced segmentation performance and reduced computational loads, respectively.
{"title":"Multi-atlas segmentation of knee cartilage via Semi-supervised Regional Label Propagation","authors":"Christos G. Chadoulos, S. Moustakidis, D. Tsaopoulos, J. Theocharis","doi":"10.1145/3507548.3507557","DOIUrl":"https://doi.org/10.1145/3507548.3507557","url":null,"abstract":"Multi-atlas based segmentation techniques have been proven to be effective in multiple automatic segmentation applications. However, mostly they rely on a non-deformable registration model followed by a voxel-wise classification process that incurs a large computational cost in terms of memory requirements and execution time. In this paper, a novel two-stage multi-atlas method is proposed, which combines constructively several concepts, including Semi-Supervised Learning (SSL), sparse graph constructions, voxel’s linear reconstructions via graph weights, and suitable sampling schemes for collecting data from target image and the atlas library. Representative global data sampled from target image are first classified according to SSL, using a newly proposed label propagation scheme. Next, out-of-sample data of yet unlabeled target voxels are iteratively generated through an iterative sampling based on mesh tetrahedralization. A thorough experimental investigation is conducted on 45 subjects provided by the publicly accessible Osteoarthritis Initiative (OAI) repository. Comparative analysis demonstrates that the proposed approach outperforms the existing state-of-the-art patch-based methods, across all evaluation metrics, exhibiting enhanced segmentation performance and reduced computational loads, respectively.","PeriodicalId":414908,"journal":{"name":"Proceedings of the 2021 5th International Conference on Computer Science and Artificial Intelligence","volume":"48 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":"114199461","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：