Micaela Jara Ten Kathen, Isabel Jurado Flores, Daniel Gutiérrez-Reina
Preserving water resources is an objective that is constantly being pursued. Monitoring the aquatic environments is an action to fulfill this objective, since the state of the water quality will be controlled. The monitoring task can be carried out with Autonomous Surface Vehicles equipped with sensors that measure water quality parameters and with a monitoring system. This paper presents a comparison between informative path planners based on PSO for autonomous surface vehicles for water resources monitoring. The case presented is the case of Ypacarai Lake. The simulations carried out allow visualizing and comparing the response of different methods. The methods evaluated are the Local Best method, the Global Best method, the Uncertainty method, the Contamination method, the Classic PSO, Enhanced GP-based PSO, and the Epsilon Greedy method. For the optimization of the Enhanced GP-based PSO coefficients, Bayesian optimization is selected. The results show that the Enhanced GP-based PSO is the algorithm with the best solutions for monitoring the lake environment.
{"title":"A Comparison of PSO-Based Informative Path Planners for Autonomous Surface Vehicles for Water Resource Monitoring","authors":"Micaela Jara Ten Kathen, Isabel Jurado Flores, Daniel Gutiérrez-Reina","doi":"10.1145/3529399.3529442","DOIUrl":"https://doi.org/10.1145/3529399.3529442","url":null,"abstract":"Preserving water resources is an objective that is constantly being pursued. Monitoring the aquatic environments is an action to fulfill this objective, since the state of the water quality will be controlled. The monitoring task can be carried out with Autonomous Surface Vehicles equipped with sensors that measure water quality parameters and with a monitoring system. This paper presents a comparison between informative path planners based on PSO for autonomous surface vehicles for water resources monitoring. The case presented is the case of Ypacarai Lake. The simulations carried out allow visualizing and comparing the response of different methods. The methods evaluated are the Local Best method, the Global Best method, the Uncertainty method, the Contamination method, the Classic PSO, Enhanced GP-based PSO, and the Epsilon Greedy method. For the optimization of the Enhanced GP-based PSO coefficients, Bayesian optimization is selected. The results show that the Enhanced GP-based PSO is the algorithm with the best solutions for monitoring the lake environment.","PeriodicalId":149111,"journal":{"name":"International Conference on Machine Learning Technologies","volume":"517 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":"116235838","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}
Nowadays, encrypted digital currency offers a new way of secure trading and exchanging and has become increasingly important in our financial system. However, the temporal dynamics of cryptocurrencies is highly complex, and predictions are still challenging. In this study, we establish two prevailing machine learning models, fully-connected Artificial Neural Network (ANN) and the Long-Short-Term-Memory (LSTM), to predictively model the price of several popular cryptocurrencies, including Bitcoin (BTC), Ethereum (ETH), Ripple (XRP), Stellar Lumens (XLM), Litecoin (LTC), and Monero (XMR). We evaluate model performance and conduct sensitivity analysis to further understand our model behaviors. We find that although LSTM seems more appropriate for time sequence prediction task, ANN, in general, outrivals LSTM in our experiments. Using price information from other different cryptocurrencies for joint training and prediction could largely facilitate the prediction of BTC. Finally, the model predictive error is highly sensitive to the time scale of interest.
{"title":"Prediction of Cryptocurrency Price Dynamics with Multiple Machine Learning Techniques","authors":"Zhengyang Wang, Xingzhou Li, Jinjin Ruan, J. Kou","doi":"10.1145/3340997.3341008","DOIUrl":"https://doi.org/10.1145/3340997.3341008","url":null,"abstract":"Nowadays, encrypted digital currency offers a new way of secure trading and exchanging and has become increasingly important in our financial system. However, the temporal dynamics of cryptocurrencies is highly complex, and predictions are still challenging. In this study, we establish two prevailing machine learning models, fully-connected Artificial Neural Network (ANN) and the Long-Short-Term-Memory (LSTM), to predictively model the price of several popular cryptocurrencies, including Bitcoin (BTC), Ethereum (ETH), Ripple (XRP), Stellar Lumens (XLM), Litecoin (LTC), and Monero (XMR). We evaluate model performance and conduct sensitivity analysis to further understand our model behaviors. We find that although LSTM seems more appropriate for time sequence prediction task, ANN, in general, outrivals LSTM in our experiments. Using price information from other different cryptocurrencies for joint training and prediction could largely facilitate the prediction of BTC. Finally, the model predictive error is highly sensitive to the time scale of interest.","PeriodicalId":149111,"journal":{"name":"International Conference on Machine Learning Technologies","volume":"39 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":"123139609","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 growing amount of available data in the biomedical domain turns out to be beneficial for decision-making, but a sufficiently accurate DR system is required. Plenty of NLP techniques and models have been proposed for semantic similarity in DR, but few of them have been able to consider the variations of the language and relationship between distant words in texts. This work is focused on formulating a Graph-based Similarity for DR method (GBS-DR) for the biomedical domain and comparing the obtained results with traditional DR paradigms. The graph-based methods were selected to prove the importance of analyzing the semantic, syntactic, and long-distant word relationships in texts. It will be demonstrated that through the graph's topology the system can extract the structural information of documents, which solves relevant issues that are faced in this research area. CCS CONCEPTS • Information Systems • Information Retrieval • Retrieval Models and Ranking • Learning to Rank
{"title":"Graph-based Similarity for Document Retrieval in the Biomedical Domain","authors":"Adelaida Zuluaga Cajiao, Andrés Rosso-Mateus","doi":"10.1145/3529399.3529428","DOIUrl":"https://doi.org/10.1145/3529399.3529428","url":null,"abstract":"The growing amount of available data in the biomedical domain turns out to be beneficial for decision-making, but a sufficiently accurate DR system is required. Plenty of NLP techniques and models have been proposed for semantic similarity in DR, but few of them have been able to consider the variations of the language and relationship between distant words in texts. This work is focused on formulating a Graph-based Similarity for DR method (GBS-DR) for the biomedical domain and comparing the obtained results with traditional DR paradigms. The graph-based methods were selected to prove the importance of analyzing the semantic, syntactic, and long-distant word relationships in texts. It will be demonstrated that through the graph's topology the system can extract the structural information of documents, which solves relevant issues that are faced in this research area. CCS CONCEPTS • Information Systems • Information Retrieval • Retrieval Models and Ranking • Learning to Rank","PeriodicalId":149111,"journal":{"name":"International Conference on Machine Learning Technologies","volume":"35 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":"121824308","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 there has been a major rise in interest in the field of semi-supervised deep learning with newer techniques being proposed which rapidly push the state-of-the-art. Most techniques, however, use balanced benchmarking datasets such as CIFAR and SVHN and therefore do not translate into real life applications, many of which involve highly imbalanced datasets. An investigation was conducted into the performance of the FixMatch algorithm when trained on unbalanced benchmarking datasets and a real world dataset. Three different entropy-based distributions of imbalance, with the proportion of labeled samples varied from 80% to 40%, were applied and compared to a baseline which was computed on uniformly balanced data. An increase in error rate is noted for the imbalanced datasets with larger errors seen in cases where there are a greater number of minority classes. Indeed, the distribution containing the most minority classes showed the maximum drop in performance with a mean error rate increase of 12.67% compared to the uniform baseline.
{"title":"Evaluating the FixMatch Semi-Supervised Algorithm for Unbalanced Image Data","authors":"A. Sajun, I. Zualkernan","doi":"10.1145/3529399.3529419","DOIUrl":"https://doi.org/10.1145/3529399.3529419","url":null,"abstract":"In recent years there has been a major rise in interest in the field of semi-supervised deep learning with newer techniques being proposed which rapidly push the state-of-the-art. Most techniques, however, use balanced benchmarking datasets such as CIFAR and SVHN and therefore do not translate into real life applications, many of which involve highly imbalanced datasets. An investigation was conducted into the performance of the FixMatch algorithm when trained on unbalanced benchmarking datasets and a real world dataset. Three different entropy-based distributions of imbalance, with the proportion of labeled samples varied from 80% to 40%, were applied and compared to a baseline which was computed on uniformly balanced data. An increase in error rate is noted for the imbalanced datasets with larger errors seen in cases where there are a greater number of minority classes. Indeed, the distribution containing the most minority classes showed the maximum drop in performance with a mean error rate increase of 12.67% compared to the uniform baseline.","PeriodicalId":149111,"journal":{"name":"International Conference on Machine Learning Technologies","volume":"23 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":"114510873","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}
Ahmed Mostafa Abdelkhalek, N. E. M. Mohamed, Mostafa M. Abdelhakam, M. M. Elmesalawy
Control measures have been applied in recent years due to the COVID-19 pandemic. Different technologies including artificial intelligence (AI) and geofencing are required to be exploited for developing efficient techniques to deal with this crisis. Workplaces are the most dangerous areas that can lead to the infection of the pandemic. This is due to the increased density of people and transactions in limited places. In this paper, an efficient approach is proposed to monitor and impose COVID-19 control measures in workplaces. The workplace environment is clustered based on a dynamic user-centric clustering scheme, where each person in the workplace is assigned to a set of associated geofences that form its cluster. For each geofence, different wireless and network metrics are used for generating its digital signature. An efficient technique based on deep learning is proposed to generate the geofence digital signature and detect whether the person is inside his associated cluster or not. Experimental results show the effectiveness of the proposed technique for different locations in a real workplace. Specifically, an accuracy of 92.86% is achieved in a workplace environment by the proposed approach.
{"title":"Dynamic User-Centric Clustered Workplaces for COVID-19 Control Measures Based on Geofencing and Deep Learning","authors":"Ahmed Mostafa Abdelkhalek, N. E. M. Mohamed, Mostafa M. Abdelhakam, M. M. Elmesalawy","doi":"10.1145/3529399.3529435","DOIUrl":"https://doi.org/10.1145/3529399.3529435","url":null,"abstract":"Control measures have been applied in recent years due to the COVID-19 pandemic. Different technologies including artificial intelligence (AI) and geofencing are required to be exploited for developing efficient techniques to deal with this crisis. Workplaces are the most dangerous areas that can lead to the infection of the pandemic. This is due to the increased density of people and transactions in limited places. In this paper, an efficient approach is proposed to monitor and impose COVID-19 control measures in workplaces. The workplace environment is clustered based on a dynamic user-centric clustering scheme, where each person in the workplace is assigned to a set of associated geofences that form its cluster. For each geofence, different wireless and network metrics are used for generating its digital signature. An efficient technique based on deep learning is proposed to generate the geofence digital signature and detect whether the person is inside his associated cluster or not. Experimental results show the effectiveness of the proposed technique for different locations in a real workplace. Specifically, an accuracy of 92.86% is achieved in a workplace environment by the proposed approach.","PeriodicalId":149111,"journal":{"name":"International Conference on Machine Learning Technologies","volume":"24 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":"121531432","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：