Pub Date : 2022-12-01DOI: 10.1109/ICMLA55696.2022.00078
A. Ranganath, Omar DeGuchy, Fabian Santiago, Mukesh Singhal, Roummel F. Marcia
Recurrent neural networks (RNNs) are traditionally used for machine learning applications for temporal sequences such as natural language processing. Its application to image processing is relatively new. In this paper, we apply RNNs to denoise images corrupted by mixed Poisson and Gaussian noise. The motivation for using an RNN comes from viewing the denoising of the Poisson-Gaussian realization as a temporal process. The network then attempts to trace back the steps that create the noisy realization in order to arrive at the noiseless reconstruction. Numerical experiments demonstrate that our proposed RNN approach outperforms convolutional autoen-coder methods for denoising and upsampling low-resolution images from the CIFAR-10 dataset.
{"title":"Recurrent Nerual Imaging: An Evolutionary Approach for Mixed Possion-Gaussian Image Denoising","authors":"A. Ranganath, Omar DeGuchy, Fabian Santiago, Mukesh Singhal, Roummel F. Marcia","doi":"10.1109/ICMLA55696.2022.00078","DOIUrl":"https://doi.org/10.1109/ICMLA55696.2022.00078","url":null,"abstract":"Recurrent neural networks (RNNs) are traditionally used for machine learning applications for temporal sequences such as natural language processing. Its application to image processing is relatively new. In this paper, we apply RNNs to denoise images corrupted by mixed Poisson and Gaussian noise. The motivation for using an RNN comes from viewing the denoising of the Poisson-Gaussian realization as a temporal process. The network then attempts to trace back the steps that create the noisy realization in order to arrive at the noiseless reconstruction. Numerical experiments demonstrate that our proposed RNN approach outperforms convolutional autoen-coder methods for denoising and upsampling low-resolution images from the CIFAR-10 dataset.","PeriodicalId":128160,"journal":{"name":"2022 21st IEEE International Conference on Machine Learning and Applications (ICMLA)","volume":"75 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133036413","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 : 2022-12-01DOI: 10.1109/ICMLA55696.2022.00257
Christoph Düsing, P. Cimiano
Federated Learning (FL) is a learning paradigm that allows clients to profit from the data that is available across multiple clients to train a joint model. As FL allows to train such a joint model without explicitly sharing data, but only sharing model updates, it has attained popularity in healthcare settings where patient data is subject to strict privacy policies and needs to be locally stored at each hospital or healthcare provider. A particular challenge for FL settings is data imbalance across clients, as it has been found to be detrimental to model performance and impact the influence of each client on the learning process. Unfortunately, the healthcare domain is particularly prone to such imbalanced data due to regional differences in disease management, prescription behavior etc. In this paper, we introduce the two novel metrics Benefit and Contribution to quantify to which degree individual clients benefit from participation in FL and how they contribute to its success, respectively. Therefore, we measure Benefit and Contribution with respect to four types of imbalances present in data at each client side. Our results show that both client Benefit and Contribution are influenced by data imbalance in such a way that high imbalance in data quantity, label distribution and feature distribution reduces or nullifies clients’ Benefit while increasing their Contribution. Thus, the most valuable clients within a cohort benefit the least from their participation, exposing a critical thread to the success of clinical FL cohorts by withdrawing participation.
{"title":"On the Trade-off Between Benefit and Contribution for Clients in Federated Learning in Healthcare","authors":"Christoph Düsing, P. Cimiano","doi":"10.1109/ICMLA55696.2022.00257","DOIUrl":"https://doi.org/10.1109/ICMLA55696.2022.00257","url":null,"abstract":"Federated Learning (FL) is a learning paradigm that allows clients to profit from the data that is available across multiple clients to train a joint model. As FL allows to train such a joint model without explicitly sharing data, but only sharing model updates, it has attained popularity in healthcare settings where patient data is subject to strict privacy policies and needs to be locally stored at each hospital or healthcare provider. A particular challenge for FL settings is data imbalance across clients, as it has been found to be detrimental to model performance and impact the influence of each client on the learning process. Unfortunately, the healthcare domain is particularly prone to such imbalanced data due to regional differences in disease management, prescription behavior etc. In this paper, we introduce the two novel metrics Benefit and Contribution to quantify to which degree individual clients benefit from participation in FL and how they contribute to its success, respectively. Therefore, we measure Benefit and Contribution with respect to four types of imbalances present in data at each client side. Our results show that both client Benefit and Contribution are influenced by data imbalance in such a way that high imbalance in data quantity, label distribution and feature distribution reduces or nullifies clients’ Benefit while increasing their Contribution. Thus, the most valuable clients within a cohort benefit the least from their participation, exposing a critical thread to the success of clinical FL cohorts by withdrawing participation.","PeriodicalId":128160,"journal":{"name":"2022 21st IEEE International Conference on Machine Learning and Applications (ICMLA)","volume":"263 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116150679","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 : 2022-12-01DOI: 10.1109/ICMLA55696.2022.00141
Ola Mustafa Alqahtani, Lakshmish Ramaswamy
In recent years, deep learning models have been widely adopted in lots of fields. such as computer vision, pattern recognition, and classification problems like plant disease classification. Due to the large diversity among the computing devices that these models may run on, we need to choose between the appropriate device based on cost and performance. Furthermore, finding the suitable optimal device for a given project is a complex process that needs significant time and resources. Prediction of inference latency DNN models is necessary for many tasks where measuring the latency on real devices is either infeasible or too costly. This is a very challenging problem, and most existing approaches fail to achieve high accuracy of prediction. While some research has been carried out to predict the inference time of DNN models – most existing techniques assume that training time is linearly related to the number of floating-point operations. This paper designs and develops a framework to predict the inference time for deep learning models and is generic to be easily extended for a large set of devices. Our key idea is decomposing a given model inference into layers and conducting layer-level prediction. Our experiments demonstrate that this strategy provides significant benefits in terms of prediction accuracy.
{"title":"A Layer Decomposition Approach to Inference Time Prediction of Deep Learning Architectures","authors":"Ola Mustafa Alqahtani, Lakshmish Ramaswamy","doi":"10.1109/ICMLA55696.2022.00141","DOIUrl":"https://doi.org/10.1109/ICMLA55696.2022.00141","url":null,"abstract":"In recent years, deep learning models have been widely adopted in lots of fields. such as computer vision, pattern recognition, and classification problems like plant disease classification. Due to the large diversity among the computing devices that these models may run on, we need to choose between the appropriate device based on cost and performance. Furthermore, finding the suitable optimal device for a given project is a complex process that needs significant time and resources. Prediction of inference latency DNN models is necessary for many tasks where measuring the latency on real devices is either infeasible or too costly. This is a very challenging problem, and most existing approaches fail to achieve high accuracy of prediction. While some research has been carried out to predict the inference time of DNN models – most existing techniques assume that training time is linearly related to the number of floating-point operations. This paper designs and develops a framework to predict the inference time for deep learning models and is generic to be easily extended for a large set of devices. Our key idea is decomposing a given model inference into layers and conducting layer-level prediction. Our experiments demonstrate that this strategy provides significant benefits in terms of prediction accuracy.","PeriodicalId":128160,"journal":{"name":"2022 21st IEEE International Conference on Machine Learning and Applications (ICMLA)","volume":"165 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116419465","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 : 2022-12-01DOI: 10.1109/ICMLA55696.2022.00083
Adrin Issai Arasu, M. Modani, N. R. Vadlamani
Temperature prediction is critical for many industrial and everyday applications. Numerical Weather Prediction (NWP) models using high-performance computing is the most sought technique to forecast weather, including temperature. However, NWP is complex in nature and computationally expensive. In this paper, the temperature is forecast using data-driven Machine Learning techniques, which are not computationally intensive and are further accelerated using GPUs. Two deep learning models: A stacked Long Short-Term Memory (LSTM) and Random Forest Regressor (RFR), are developed and validated using the standard ERA5 data (at 850hPa, above the atmospheric boundary layer). In addition, the models are tested against the ground-level observations (inside the atmospheric boundary layer) for twenty different locations in India. The performance of univariate and multivariate models is also analyzed for the real-time dataset. Root Mean Square Error (RMSE) obtained by the LSTM and RFR are 0.47 and 0.23, respectively, for ERA5 data. When compared to the numerical weather prediction model - operational IFS, the RMSE using LSTM and RFR is smaller by 65% and 83%, respectively. The LSTM and RFR models forecast temperature with an average RMSE of 0.7 for the real-time data at twenty locations. The GPU-enabled LSTM model performed 64 times faster than the CPU-enabled model. The developed RNN models are made publicly available at https://github.com/arasuadrian/RNN-Models.
{"title":"Application of Machine Learning Techniques in Temperature Forecast","authors":"Adrin Issai Arasu, M. Modani, N. R. Vadlamani","doi":"10.1109/ICMLA55696.2022.00083","DOIUrl":"https://doi.org/10.1109/ICMLA55696.2022.00083","url":null,"abstract":"Temperature prediction is critical for many industrial and everyday applications. Numerical Weather Prediction (NWP) models using high-performance computing is the most sought technique to forecast weather, including temperature. However, NWP is complex in nature and computationally expensive. In this paper, the temperature is forecast using data-driven Machine Learning techniques, which are not computationally intensive and are further accelerated using GPUs. Two deep learning models: A stacked Long Short-Term Memory (LSTM) and Random Forest Regressor (RFR), are developed and validated using the standard ERA5 data (at 850hPa, above the atmospheric boundary layer). In addition, the models are tested against the ground-level observations (inside the atmospheric boundary layer) for twenty different locations in India. The performance of univariate and multivariate models is also analyzed for the real-time dataset. Root Mean Square Error (RMSE) obtained by the LSTM and RFR are 0.47 and 0.23, respectively, for ERA5 data. When compared to the numerical weather prediction model - operational IFS, the RMSE using LSTM and RFR is smaller by 65% and 83%, respectively. The LSTM and RFR models forecast temperature with an average RMSE of 0.7 for the real-time data at twenty locations. The GPU-enabled LSTM model performed 64 times faster than the CPU-enabled model. The developed RNN models are made publicly available at https://github.com/arasuadrian/RNN-Models.","PeriodicalId":128160,"journal":{"name":"2022 21st IEEE International Conference on Machine Learning and Applications (ICMLA)","volume":"29 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121776108","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 : 2022-12-01DOI: 10.1109/ICMLA55696.2022.00122
Daniel Lauss, F. Eibensteiner, P. Petz
In this paper, we demonstrate an inertial measurement unit (IMU) based hand gesture recognition (HGR) on a low-power microcontroller (STM32L476JGY). The focus of this work is to build a reliable hardware prototype by using deep neural networks (DNN) deployed on a resource limited device. To train the DNNs, a dataset was recorded which contains accelerometer and gyroscope readings from three IMUs mounted on the fingertips. With this dataset, various neural networks (NN) were trained and analyzed. The best NN, in terms of accuracy, memory usage and latency, was then selected and ported to the microcontroller. Finally, a runtime analysis of the model has been performed on the controller. The analysis showed that a LSTM is best suited for the detection of hand gestures. The selected model achieves an accuracy of 93% and only takes up around 40KiB of memory. In addition, the model has a throughput time of only 3.52ms, which means that the prototype can be used in real time.
{"title":"A Deep Learning based Hand Gesture Recognition on a Low-power Microcontroller using IMU Sensors","authors":"Daniel Lauss, F. Eibensteiner, P. Petz","doi":"10.1109/ICMLA55696.2022.00122","DOIUrl":"https://doi.org/10.1109/ICMLA55696.2022.00122","url":null,"abstract":"In this paper, we demonstrate an inertial measurement unit (IMU) based hand gesture recognition (HGR) on a low-power microcontroller (STM32L476JGY). The focus of this work is to build a reliable hardware prototype by using deep neural networks (DNN) deployed on a resource limited device. To train the DNNs, a dataset was recorded which contains accelerometer and gyroscope readings from three IMUs mounted on the fingertips. With this dataset, various neural networks (NN) were trained and analyzed. The best NN, in terms of accuracy, memory usage and latency, was then selected and ported to the microcontroller. Finally, a runtime analysis of the model has been performed on the controller. The analysis showed that a LSTM is best suited for the detection of hand gestures. The selected model achieves an accuracy of 93% and only takes up around 40KiB of memory. In addition, the model has a throughput time of only 3.52ms, which means that the prototype can be used in real time.","PeriodicalId":128160,"journal":{"name":"2022 21st IEEE International Conference on Machine Learning and Applications (ICMLA)","volume":"46 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125983775","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 : 2022-12-01DOI: 10.1109/ICMLA55696.2022.00115
Noel Elias
Modern day audio signal classification techniques lack the ability to classify low feature audio signals in the form of spectrographic temporal frequency data representations. Additionally, currently utilized techniques rely on full diverse data sets that are often not representative of real-world distributions. This paper derives several first-of-its-kind machine learning methodologies to analyze these low feature audio spectrograms given data distributions that may have normalized, skewed, or even limited training sets. In particular, this paper proposes several novel customized convolutional architectures to extract identifying features using binary, one-class, and siamese approaches to identify the spectrographic signature of a given audio signal. Utilizing these novel convolutional architectures as well as the proposed classification methods, these experiments demonstrate state-of-the-art classification accuracy and improved efficiency than traditional audio classification methods.
{"title":"Audio Classification of Low Feature Spectrograms Utilizing Convolutional Neural Networks","authors":"Noel Elias","doi":"10.1109/ICMLA55696.2022.00115","DOIUrl":"https://doi.org/10.1109/ICMLA55696.2022.00115","url":null,"abstract":"Modern day audio signal classification techniques lack the ability to classify low feature audio signals in the form of spectrographic temporal frequency data representations. Additionally, currently utilized techniques rely on full diverse data sets that are often not representative of real-world distributions. This paper derives several first-of-its-kind machine learning methodologies to analyze these low feature audio spectrograms given data distributions that may have normalized, skewed, or even limited training sets. In particular, this paper proposes several novel customized convolutional architectures to extract identifying features using binary, one-class, and siamese approaches to identify the spectrographic signature of a given audio signal. Utilizing these novel convolutional architectures as well as the proposed classification methods, these experiments demonstrate state-of-the-art classification accuracy and improved efficiency than traditional audio classification methods.","PeriodicalId":128160,"journal":{"name":"2022 21st IEEE International Conference on Machine Learning and Applications (ICMLA)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125490721","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 : 2022-12-01DOI: 10.1109/ICMLA55696.2022.00084
W. Rocha, Antônio U Lucena, G. F. Sarmanho, Rodrigo C Félix, S. Miqueleti, T. C. Dourado
Dam monitoring can be used as an important indicator for dam risk management. In this study, a methodology based on machine learning and ultrasound for dam safety monitoring is presented. First, a prototype dam was built to simulate different environmental conditions. Second, ultrasound images were acquired in different areas of a prototype dam. Finally, various machine learning algorithms were applied to distinguish the different regions observed in the prototype dam. The results show that it is possible to distinguish the dam regions, which is of great value for dam safety monitoring and operation.
{"title":"Machine learning protocol from ultrasound data for monitoring, predicting, and supporting the analysis of dam slopes","authors":"W. Rocha, Antônio U Lucena, G. F. Sarmanho, Rodrigo C Félix, S. Miqueleti, T. C. Dourado","doi":"10.1109/ICMLA55696.2022.00084","DOIUrl":"https://doi.org/10.1109/ICMLA55696.2022.00084","url":null,"abstract":"Dam monitoring can be used as an important indicator for dam risk management. In this study, a methodology based on machine learning and ultrasound for dam safety monitoring is presented. First, a prototype dam was built to simulate different environmental conditions. Second, ultrasound images were acquired in different areas of a prototype dam. Finally, various machine learning algorithms were applied to distinguish the different regions observed in the prototype dam. The results show that it is possible to distinguish the dam regions, which is of great value for dam safety monitoring and operation.","PeriodicalId":128160,"journal":{"name":"2022 21st IEEE International Conference on Machine Learning and Applications (ICMLA)","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125897673","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 : 2022-12-01DOI: 10.1109/ICMLA55696.2022.00264
Igor Matias, K. Wac
Wearable devices are a useful and widely used source of continuous and temporal dependant data. In contrast to the traditional clinical environment, these devices allow time series data collection in an individual’s daily living environment. However, missing data can occur while using them. Many techniques have been applied to solve these data gaps; nonetheless, missing time series data poses extra challenges, such as maintaining the temporal dependency. In this article, we addressed the forecast of sleep trackers data (sleeping heart rate (HR) and time asleep) for 2 main reasons: (1) to design models capable of accurately forecasting missing data from those devices, and (2) to apply those models to empower sleep interventions that may increase its quality, by forecasting future sleep events. We collected wearables data over 290 days (per individual) from 12 participants using a smartwatch and made this dataset publicly available. We then explored several hyperparameters of 2 Recurrent Neural Networks (RNN), Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU). We further elaborated and compared the performance of 3 approaches to training those RNNs. Although similar performance, slightly more accurate results were obtained after training a GRU network on an entire population’s dataset, which was able to forecast the average, minimum, and maximum sleeping HR with a root-mean-squared error (RMSE) of 4.4 (± 1.4), 4.9 (± 2.6), and 12.1 ( 4.0) beats per minute, respectively. However, the total time ±asleep was impossible to forecast with low error.
{"title":"Dejà vu: Recurrent Neural Networks for health wearables data forecast","authors":"Igor Matias, K. Wac","doi":"10.1109/ICMLA55696.2022.00264","DOIUrl":"https://doi.org/10.1109/ICMLA55696.2022.00264","url":null,"abstract":"Wearable devices are a useful and widely used source of continuous and temporal dependant data. In contrast to the traditional clinical environment, these devices allow time series data collection in an individual’s daily living environment. However, missing data can occur while using them. Many techniques have been applied to solve these data gaps; nonetheless, missing time series data poses extra challenges, such as maintaining the temporal dependency. In this article, we addressed the forecast of sleep trackers data (sleeping heart rate (HR) and time asleep) for 2 main reasons: (1) to design models capable of accurately forecasting missing data from those devices, and (2) to apply those models to empower sleep interventions that may increase its quality, by forecasting future sleep events. We collected wearables data over 290 days (per individual) from 12 participants using a smartwatch and made this dataset publicly available. We then explored several hyperparameters of 2 Recurrent Neural Networks (RNN), Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU). We further elaborated and compared the performance of 3 approaches to training those RNNs. Although similar performance, slightly more accurate results were obtained after training a GRU network on an entire population’s dataset, which was able to forecast the average, minimum, and maximum sleeping HR with a root-mean-squared error (RMSE) of 4.4 (± 1.4), 4.9 (± 2.6), and 12.1 ( 4.0) beats per minute, respectively. However, the total time ±asleep was impossible to forecast with low error.","PeriodicalId":128160,"journal":{"name":"2022 21st IEEE International Conference on Machine Learning and Applications (ICMLA)","volume":"10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130282337","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 : 2022-12-01DOI: 10.1109/ICMLA55696.2022.00164
Jerry Bonnell, Melanie Xia, Lee Wall, York Eggleston, M. Ogihara, V. Aguiar-Pulido
Personalized medicine is becoming of increasing importance in the study of psoriasis and atopic dermatitis (AD). Because current treatments only target symptoms, early intervention and personalized medicine have a pivotal role in improved health outcomes. To explore this potential, this study investigates the use of direct-to-consumer (DTC) genetic data in devising machine learning models that can pinpoint signatures salient to psoriasis and AD. The study simulates high-dimensional datasets derived from the HapMap 3 and 1000 Genomes Project cohorts (561K and 497K loci, respectively, that act as features). The simulation scheme splits subjects into cases and controls, where randomly selected variants associated with the target phenotypes are introduced into the cases. Unsupervised learning (UMAP) and eight supervised learning techniques are applied to each of the simulated datasets. Our findings suggest that the parametric models tested (SVM, LASSO, and RIDGE) exhibit the best predictive power on the simulated datasets while also yielding high retrieval rates for signatures associated with the target phenotypes.
{"title":"Machine Learning in Personalized Skin Care: A Simulation Scheme for Pattern Recognition in Skin Condition Genome-wide Association Studies","authors":"Jerry Bonnell, Melanie Xia, Lee Wall, York Eggleston, M. Ogihara, V. Aguiar-Pulido","doi":"10.1109/ICMLA55696.2022.00164","DOIUrl":"https://doi.org/10.1109/ICMLA55696.2022.00164","url":null,"abstract":"Personalized medicine is becoming of increasing importance in the study of psoriasis and atopic dermatitis (AD). Because current treatments only target symptoms, early intervention and personalized medicine have a pivotal role in improved health outcomes. To explore this potential, this study investigates the use of direct-to-consumer (DTC) genetic data in devising machine learning models that can pinpoint signatures salient to psoriasis and AD. The study simulates high-dimensional datasets derived from the HapMap 3 and 1000 Genomes Project cohorts (561K and 497K loci, respectively, that act as features). The simulation scheme splits subjects into cases and controls, where randomly selected variants associated with the target phenotypes are introduced into the cases. Unsupervised learning (UMAP) and eight supervised learning techniques are applied to each of the simulated datasets. Our findings suggest that the parametric models tested (SVM, LASSO, and RIDGE) exhibit the best predictive power on the simulated datasets while also yielding high retrieval rates for signatures associated with the target phenotypes.","PeriodicalId":128160,"journal":{"name":"2022 21st IEEE International Conference on Machine Learning and Applications (ICMLA)","volume":"35 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126363065","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 : 2022-12-01DOI: 10.1109/ICMLA55696.2022.00081
S. Sharifzadeh
X-ray Computed Tomography (CT) imaging is used as a non-destructive strategy for characterizing the internal structure of rocks. One important application of such studies is prediction of the relative permeability of CO2 in reservoirs. Estimation of Carbon Capture and Storage (CCS) has a great impact in mitigation strategies for global warming and controlling the effects of climate change. In this paper, 3D Xray Computed Tomography (CT) image volumes of rocks are characterized for prediction of the CO2 relative permeability. A new analysis pipeline is introduced that extracts high dimensional entropy features from the local 3D voxels. That is followed by a sparse kernelized dimensionality reduction step to alleviate the over-fitting issue. Then, regression analysis is performed using Gaussian Process Regression (GPR). Furthermore, the proposed pipeline is compared with two other deep Neural Networks (NN) models including a Convolutional Neural Network (CNN) regression model as well as a transferred pre-trained ResNet50 model using the rock X-ray training data. Experimental results show improvements in CO2 permeability prediction using the proposed analysis pipeline.
{"title":"Spars Kernelized Features for Prediction of Rock’s Carbon Capture using 3D X-Ray Images","authors":"S. Sharifzadeh","doi":"10.1109/ICMLA55696.2022.00081","DOIUrl":"https://doi.org/10.1109/ICMLA55696.2022.00081","url":null,"abstract":"X-ray Computed Tomography (CT) imaging is used as a non-destructive strategy for characterizing the internal structure of rocks. One important application of such studies is prediction of the relative permeability of CO2 in reservoirs. Estimation of Carbon Capture and Storage (CCS) has a great impact in mitigation strategies for global warming and controlling the effects of climate change. In this paper, 3D Xray Computed Tomography (CT) image volumes of rocks are characterized for prediction of the CO2 relative permeability. A new analysis pipeline is introduced that extracts high dimensional entropy features from the local 3D voxels. That is followed by a sparse kernelized dimensionality reduction step to alleviate the over-fitting issue. Then, regression analysis is performed using Gaussian Process Regression (GPR). Furthermore, the proposed pipeline is compared with two other deep Neural Networks (NN) models including a Convolutional Neural Network (CNN) regression model as well as a transferred pre-trained ResNet50 model using the rock X-ray training data. Experimental results show improvements in CO2 permeability prediction using the proposed analysis pipeline.","PeriodicalId":128160,"journal":{"name":"2022 21st IEEE International Conference on Machine Learning and Applications (ICMLA)","volume":"86 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126354542","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：