Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
4.7
2.9
Journals
Electronics
Special Issues
Advances in Machine Learning
share announcement

Advances in Machine Learning

A special issue of Electronics (ISSN 2079-9292). This special issue belongs to the section "Computer Science & Engineering".

Deadline for manuscript submissions: closed (31 January 2022) | Viewed by 72793

Special Issue Editors

Prof. Dr. Jihoon Yang

E-Mail Website
Guest Editor
Department of Computer Science and Engineering, Sogang University, Seoul 04107, Republic of Korea
Interests: machine learning; artificial intelligence; data mining
Special Issues, Collections and Topics in MDPI journals
Dr. Unsang Park

E-Mail Website
Guest Editor
Department of Computer Science and Engineering, Sogang University, Seoul 04107, Korea
Interests: computer vision; pattern recognition; biometrics
Special Issues, Collections and Topics in MDPI journals

Special Issue Information

Dear Colleagues,  

Today, machine learning which aims to teach computers in a bid to make them act like human has become essential. A number of algorithms, techniques, and methodologies have been proposed for a variety of tasks, including autonomous driving, game playing, disease diagnosis and treatment, fraud detection, spam filtering, speech recognition, object detection, search, and recommendation.  

This Special Issue is seeking high-quality research papers in all areas of machine learning. It is open to well-organized reviews as well as application papers. Topics include but are not limited to the following: 

  • Deep learning;
  • Reinforcement learning;
  • Automated machine learning;
  • On-device learning;
  • Transfer learning;
  • Meta learning;
  • Application of machine learning in real-world domains. 

Prof. Dr. Jihoon Yang
Prof. Dr. Unsang Park
Guest Editors

Manuscript Submission Information

Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 100 words) can be sent to the Editorial Office for announcement on this website.

Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Electronics is an international peer-reviewed open access semimonthly journal published by MDPI.

Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 2400 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.

Keywords

  • machine learning
  • deep learning
  • data mining and analysis

Published Papers (17 papers)

Download All Papers
Order results
Result details
Show export options expand_more Show export options expand_less
Select all
Export citation of selected articles as:

Editorial

Jump to: Research

3 pages, 172 KiB  
Editorial
Advances in Machine Learning
by Jihoon Yang and Unsang Park
Electronics 2022, 11(9), 1428; https://doi.org/10.3390/electronics11091428 - 29 Apr 2022
Viewed by 1265
Abstract
Since its inception as a branch of Artificial Intelligence, Machine Learning (ML) has flourished in recent years [...] Full article
(This article belongs to the Special Issue Advances in Machine Learning)

Research

Jump to: Editorial

13 pages, 2596 KiB  
Article
A Deep Learning Model for Network Intrusion Detection with Imbalanced Data
by Yanfang Fu, Yishuai Du, Zijian Cao, Qiang Li and Wei Xiang
Electronics 2022, 11(6), 898; https://doi.org/10.3390/electronics11060898 - 14 Mar 2022
Cited by 91 | Viewed by 7040
Abstract
With an increase in the number and types of network attacks, traditional firewalls and data encryption methods can no longer meet the needs of current network security. As a result, intrusion detection systems have been proposed to deal with network threats. The current [...] Read more.
With an increase in the number and types of network attacks, traditional firewalls and data encryption methods can no longer meet the needs of current network security. As a result, intrusion detection systems have been proposed to deal with network threats. The current mainstream intrusion detection algorithms are aided with machine learning but have problems of low detection rates and the need for extensive feature engineering. To address the issue of low detection accuracy, this paper proposes a model for traffic anomaly detection named a deep learning model for network intrusion detection (DLNID), which combines an attention mechanism and the bidirectional long short-term memory (Bi-LSTM) network, first extracting sequence features of data traffic through a convolutional neural network (CNN) network, then reassigning the weights of each channel through the attention mechanism, and finally using Bi-LSTM to learn the network of sequence features. In intrusion detection public data sets, there are serious imbalance data generally. To address data imbalance issues, this paper employs the method of adaptive synthetic sampling (ADASYN) for sample expansion of minority class samples, to eventually form a relatively symmetric dataset, and uses a modified stacked autoencoder for data dimensionality reduction with the objective of enhancing information fusion. DLNID is an end-to-end model, so it does not need to undergo the process of manual feature extraction. After being tested on the public benchmark dataset on network intrusion detection NSL-KDD, experimental results show that the accuracy and F1 score of this model are better than those of other comparison methods, reaching 90.73% and 89.65%, respectively. Full article
(This article belongs to the Special Issue Advances in Machine Learning)
►▼ Show Figures

Graphical abstract

10 pages, 5781 KiB  
Article
Predictive Capacity Planning for Mobile Networks—ML Supported Prediction of Network Performance and User Experience Evolution
by Igor Tomic, Eoin Bleakley and Predrag Ivanis
Electronics 2022, 11(4), 626; https://doi.org/10.3390/electronics11040626 - 17 Feb 2022
Cited by 6 | Viewed by 3274
Abstract
Network performance prediction is crucial for enabling agile capacity planning in mobile networks. One of the key problems is predicting evolution of spectral efficiency in growing network load conditions. The main factor driving network performance and spectral efficiency is reportedly the Channel Quality [...] Read more.
Network performance prediction is crucial for enabling agile capacity planning in mobile networks. One of the key problems is predicting evolution of spectral efficiency in growing network load conditions. The main factor driving network performance and spectral efficiency is reportedly the Channel Quality Indicator (CQI). In this paper, the performance of different Machine Learning (ML) models were examined, and XGBoost was selected as the best performing model. Furthermore, to improve modeling accuracy, several features were introduced (operating frequency band, Physical Resource Block (PRB) utilization in surrounding cells, number of surrounding cells within a radius, heavy data factor and higher order modulation usage). The impact of these features on CQI prediction were examined. Full article
(This article belongs to the Special Issue Advances in Machine Learning)
►▼ Show Figures

Figure 1

17 pages, 2106 KiB  
Article
Blind Image Super Resolution Using Deep Unsupervised Learning
by Kazuhiro Yamawaki, Yongqing Sun and Xian-Hua Han
Electronics 2021, 10(21), 2591; https://doi.org/10.3390/electronics10212591 - 23 Oct 2021
Cited by 4 | Viewed by 2279
Abstract
The goal of single image super resolution (SISR) is to recover a high-resolution (HR) image from a low-resolution (LR) image. Deep learning based methods have recently made a remarkable performance gain in terms of both the effectiveness and efficiency for SISR. Most existing [...] Read more.
The goal of single image super resolution (SISR) is to recover a high-resolution (HR) image from a low-resolution (LR) image. Deep learning based methods have recently made a remarkable performance gain in terms of both the effectiveness and efficiency for SISR. Most existing methods have to be trained based on large-scale synthetic paired data in a fully supervised manner. With the available HR natural images, the corresponding LR images are usually synthesized with a simple fixed degradation operation, such as bicubic down-sampling. Then, the learned deep models with these training data usually face difficulty to be generalized to real scenarios with unknown and complicated degradation operations. This study exploits a novel blind image super-resolution framework using a deep unsupervised learning network. The proposed method can simultaneously predict the underlying HR image and its specific degradation operation from the observed LR image only without any prior knowledge. The experimental results on three benchmark datasets validate that our proposed method achieves a promising performance under the unknown degradation models. Full article
(This article belongs to the Special Issue Advances in Machine Learning)
►▼ Show Figures

Figure 1

15 pages, 350 KiB  
Article
Unsupervised Outlier Detection: A Meta-Learning Algorithm Based on Feature Selection
by Vasilis Papastefanopoulos, Pantelis Linardatos and Sotiris Kotsiantis
Electronics 2021, 10(18), 2236; https://doi.org/10.3390/electronics10182236 - 12 Sep 2021
Cited by 5 | Viewed by 3560
Abstract
Outlier detection refers to the problem of the identification and, where appropriate, the elimination of anomalous observations from data. Such anomalous observations can emerge due to a variety of reasons, including human or mechanical errors, fraudulent behaviour as well as environmental or systematic [...] Read more.
Outlier detection refers to the problem of the identification and, where appropriate, the elimination of anomalous observations from data. Such anomalous observations can emerge due to a variety of reasons, including human or mechanical errors, fraudulent behaviour as well as environmental or systematic changes, occurring either naturally or purposefully. The accurate and timely detection of deviant observations allows for the early identification of potentially extensive problems, such as fraud or system failures, before they escalate. Several unsupervised outlier detection methods have been developed; however, there is no single best algorithm or family of algorithms, as typically each relies on a measure of ‘outlierness’ such as density or distance, ignoring other measures. To add to that, in an unsupervised setting, the absence of ground-truth labels makes finding a single best algorithm an impossible feat even for a single given dataset. In this study, a new meta-learning algorithm for unsupervised outlier detection is introduced in order to mitigate this problem. The proposed algorithm, in a fully unsupervised manner, attempts not only to combine the best of many worlds from the existing techniques through ensemble voting but also mitigate any undesired shortcomings by employing an unsupervised feature selection strategy in order to identify the most informative algorithms for a given dataset. The proposed methodology was evaluated extensively through experimentation, where it was benchmarked and compared against a wide range of commonly-used techniques for outlier detection. Results obtained using a variety of widely accepted datasets demonstrated its usefulness and its state-of-the-art results as it topped the Friedman ranking test for both the area under receiver operating characteristic (ROC) curve and precision metrics when averaged over five independent trials. Full article
(This article belongs to the Special Issue Advances in Machine Learning)
►▼ Show Figures

Figure 1

27 pages, 2842 KiB  
Article
Multi-Level Deep Learning Model for Potato Leaf Disease Recognition
by Javed Rashid, Imran Khan, Ghulam Ali, Sultan H. Almotiri, Mohammed A. AlGhamdi and Khalid Masood
Electronics 2021, 10(17), 2064; https://doi.org/10.3390/electronics10172064 - 26 Aug 2021
Cited by 78 | Viewed by 12257
Abstract
Potato leaf disease detection in an early stage is challenging because of variations in crop species, crop diseases symptoms and environmental factors. These factors make it difficult to detect potato leaf diseases in the early stage. Various machine learning techniques have been developed [...] Read more.
Potato leaf disease detection in an early stage is challenging because of variations in crop species, crop diseases symptoms and environmental factors. These factors make it difficult to detect potato leaf diseases in the early stage. Various machine learning techniques have been developed to detect potato leaf diseases. However, the existing methods cannot detect crop species and crop diseases in general because these models are trained and tested on images of plant leaves of a specific region. In this research, a multi-level deep learning model for potato leaf disease recognition has developed. At the first level, it extracts the potato leaves from the potato plant image using the YOLOv5 image segmentation technique. At the second level, a novel deep learning technique has been developed using a convolutional neural network to detect the early blight and late blight potato diseases from potato leaf images. The proposed potato leaf disease detection model was trained and tested on a potato leaf disease dataset. The potato leaf disease dataset contains 4062 images collected from the Central Punjab region of Pakistan. The proposed deep learning technique achieved 99.75% accuracy on the potato leaf disease dataset. The performance of the proposed techniques was also evaluated on the PlantVillage dataset. The proposed technique is also compared with the state-of-the-art models and achieved significantly concerning the accuracy and computational cost. Full article
(This article belongs to the Special Issue Advances in Machine Learning)
►▼ Show Figures

Figure 1

25 pages, 2988 KiB  
Article
Multi-Path Deep CNN with Residual Inception Network for Single Image Super-Resolution
by Wazir Muhammad, Zuhaibuddin Bhutto, Arslan Ansari, Mudasar Latif Memon, Ramesh Kumar, Ayaz Hussain, Syed Ali Raza Shah, Imdadullah Thaheem and Shamshad Ali
Electronics 2021, 10(16), 1979; https://doi.org/10.3390/electronics10161979 - 17 Aug 2021
Cited by 11 | Viewed by 3545
Abstract
Recent research on single-image super-resolution (SISR) using deep convolutional neural networks has made a breakthrough and achieved tremendous performance. Despite their significant progress, numerous convolutional neural networks (CNN) are limited in practical applications, owing to the requirement of the heavy computational cost of [...] Read more.
Recent research on single-image super-resolution (SISR) using deep convolutional neural networks has made a breakthrough and achieved tremendous performance. Despite their significant progress, numerous convolutional neural networks (CNN) are limited in practical applications, owing to the requirement of the heavy computational cost of the model. This paper proposes a multi-path network for SISR, known as multi-path deep CNN with residual inception network for single image super-resolution. In detail, a residual/ResNet block with an Inception block supports the main framework of the entire network architecture. In addition, remove the batch normalization layer from the residual network (ResNet) block and max-pooling layer from the Inception block to further reduce the number of parameters to preventing the over-fitting problem during the training. Moreover, a conventional rectified linear unit (ReLU) is replaced with Leaky ReLU activation function to speed up the training process. Specifically, we propose a novel two upscale module, which adopts three paths to upscale the features by jointly using deconvolution and upsampling layers, instead of using single deconvolution layer or upsampling layer alone. The extensive experimental results on image super-resolution (SR) using five publicly available test datasets, which show that the proposed model not only attains the higher score of peak signal-to-noise ratio/structural similarity index matrix (PSNR/SSIM) but also enables faster and more efficient calculations against the existing image SR methods. For instance, we improved our method in terms of overall PSNR on the SET5 dataset with challenging upscale factor 8× as 1.88 dB over the baseline bicubic method and reduced computational cost in terms of number of parameters 62% by deeply-recursive convolutional neural network (DRCN) method. Full article
(This article belongs to the Special Issue Advances in Machine Learning)
►▼ Show Figures

Figure 1

23 pages, 3688 KiB  
Article
Greed Is Good: Rapid Hyperparameter Optimization and Model Selection Using Greedy k-Fold Cross Validation
by Daniel S. Soper
Electronics 2021, 10(16), 1973; https://doi.org/10.3390/electronics10161973 - 16 Aug 2021
Cited by 21 | Viewed by 3785
Abstract
Selecting a final machine learning (ML) model typically occurs after a process of hyperparameter optimization in which many candidate models with varying structural properties and algorithmic settings are evaluated and compared. Evaluating each candidate model commonly relies on k-fold cross validation, wherein [...] Read more.
Selecting a final machine learning (ML) model typically occurs after a process of hyperparameter optimization in which many candidate models with varying structural properties and algorithmic settings are evaluated and compared. Evaluating each candidate model commonly relies on k-fold cross validation, wherein the data are randomly subdivided into k folds, with each fold being iteratively used as a validation set for a model that has been trained using the remaining folds. While many research studies have sought to accelerate ML model selection by applying metaheuristic and other search methods to the hyperparameter space, no consideration has been given to the k-fold cross validation process itself as a means of rapidly identifying the best-performing model. The current study rectifies this oversight by introducing a greedy k-fold cross validation method and demonstrating that greedy k-fold cross validation can vastly reduce the average time required to identify the best-performing model when given a fixed computational budget and a set of candidate models. This improved search time is shown to hold across a variety of ML algorithms and real-world datasets. For scenarios without a computational budget, this paper also introduces an early stopping algorithm based on the greedy cross validation method. The greedy early stopping method is shown to outperform a competing, state-of-the-art early stopping method both in terms of search time and the quality of the ML models selected by the algorithm. Since hyperparameter optimization is among the most time-consuming, computationally intensive, and monetarily expensive tasks in the broader process of developing ML-based solutions, the ability to rapidly identify optimal machine learning models using greedy cross validation has obvious and substantial benefits to organizations and researchers alike. Full article
(This article belongs to the Special Issue Advances in Machine Learning)
►▼ Show Figures

Figure 1

17 pages, 20924 KiB  
Article
RMU-Net: A Novel Residual Mobile U-Net Model for Brain Tumor Segmentation from MR Images
by Muhammad Usman Saeed, Ghulam Ali, Wang Bin, Sultan H. Almotiri, Mohammed A. AlGhamdi, Arfan Ali Nagra, Khalid Masood and Riaz ul Amin
Electronics 2021, 10(16), 1962; https://doi.org/10.3390/electronics10161962 - 14 Aug 2021
Cited by 30 | Viewed by 6155
Abstract
The most aggressive form of brain tumor is gliomas, which leads to concise life when high grade. The early detection of glioma is important to save the life of patients. MRI is a commonly used approach for brain tumors evaluation. However, the massive [...] Read more.
The most aggressive form of brain tumor is gliomas, which leads to concise life when high grade. The early detection of glioma is important to save the life of patients. MRI is a commonly used approach for brain tumors evaluation. However, the massive amount of data provided by MRI prevents manual segmentation in a reasonable time, restricting the use of accurate quantitative measurements in clinical practice. An automatic and reliable method is required that can segment tumors accurately. To achieve end-to-end brain tumor segmentation, a hybrid deep learning model RMU-Net is proposed. The architecture of MobileNetV2 is modified by adding residual blocks to learn in-depth features. This modified Mobile Net V2 is used as an encoder in the proposed network, and upsampling layers of U-Net are used as the decoder part. The proposed model has been validated on BraTS 2020, BraTS 2019, and BraTS 2018 datasets. The RMU-Net achieved the dice coefficient scores for WT, TC, and ET of 91.35%, 88.13%, and 83.26% on the BraTS 2020 dataset, 91.76%, 91.23%, and 83.19% on the BraTS 2019 dataset, and 90.80%, 86.75%, and 79.36% on the BraTS 2018 dataset, respectively. The performance of the proposed method outperforms with less computational cost and time as compared to previous methods. Full article
(This article belongs to the Special Issue Advances in Machine Learning)
►▼ Show Figures

Figure 1

13 pages, 264 KiB  
Article
Decision Tree Application to Classification Problems with Boosting Algorithm
by Long Zhao, Sanghyuk Lee and Seon-Phil Jeong
Electronics 2021, 10(16), 1903; https://doi.org/10.3390/electronics10161903 - 8 Aug 2021
Cited by 28 | Viewed by 3257
Abstract
A personal credit evaluation algorithm is proposed by the design of a decision tree with a boosting algorithm, and the classification is carried out. By comparison with the conventional decision tree algorithm, it is shown that the boosting algorithm acts to speed up [...] Read more.
A personal credit evaluation algorithm is proposed by the design of a decision tree with a boosting algorithm, and the classification is carried out. By comparison with the conventional decision tree algorithm, it is shown that the boosting algorithm acts to speed up the processing time. The Classification and Regression Tree (CART) algorithm with the boosting algorithm showed 90.95% accuracy, slightly higher than without boosting, 90.31%. To avoid overfitting of the model on the training set due to unreasonable data set division, we consider cross-validation and illustrate the results with simulation; hypermeters of the model have been applied and the model fitting effect is verified. The proposed decision tree model is fitted optimally with the help of a confusion matrix. In this paper, relevant evaluation indicators are also introduced to evaluate the performance of the proposed model. For the comparison with the conventional methods, accuracy rate, error rate, precision, recall, etc. are also illustrated; we comprehensively evaluate the model performance based on the model accuracy after the 10-fold cross-validation. The results show that the boosting algorithm improves the performance of the model in accuracy and precision when CART is applied, but the model fitting time takes much longer, around 2 min. With the obtained result, it is verified that the performance of the decision tree model is improved under the boosting algorithm. At the same time, we test the performance of the proposed verification model with model fitting, and it could be applied to the prediction model for customers’ decisions on subscription to the fixed deposit business. Full article
(This article belongs to the Special Issue Advances in Machine Learning)
13 pages, 1078 KiB  
Article
Progressive Convolutional Neural Network for Incremental Learning
by Zahid Ali Siddiqui and Unsang Park
Electronics 2021, 10(16), 1879; https://doi.org/10.3390/electronics10161879 - 5 Aug 2021
Cited by 6 | Viewed by 2884
Abstract
In this paper, we present a novel incremental learning technique to solve the catastrophic forgetting problem observed in the CNN architectures. We used a progressive deep neural network to incrementally learn new classes while keeping the performance of the network unchanged on old [...] Read more.
In this paper, we present a novel incremental learning technique to solve the catastrophic forgetting problem observed in the CNN architectures. We used a progressive deep neural network to incrementally learn new classes while keeping the performance of the network unchanged on old classes. The incremental training requires us to train the network only for new classes and fine-tune the final fully connected layer, without needing to train the entire network again, which significantly reduces the training time. We evaluate the proposed architecture extensively on image classification task using Fashion MNIST, CIFAR-100 and ImageNet-1000 datasets. Experimental results show that the proposed network architecture not only alleviates catastrophic forgetting but can also leverages prior knowledge via lateral connections to previously learned classes and their features. In addition, the proposed scheme is easily scalable and does not require structural changes on the network trained on the old task, which are highly required properties in embedded systems. Full article
(This article belongs to the Special Issue Advances in Machine Learning)
►▼ Show Figures

Figure 1

18 pages, 2746 KiB  
Article
Automatic Classification of Monosyllabic and Multisyllabic Birds Using PDHF
by Abdullah Alghamdi, Tooba Mehtab, Rizwan Iqbal, Mona Leeza, Noman Islam, Mohammed Hamdi and Asadullah Shaikh
Electronics 2021, 10(5), 624; https://doi.org/10.3390/electronics10050624 - 8 Mar 2021
Cited by 5 | Viewed by 2097
Abstract
Bioacoustics plays an important role in the conservation of bird species. Bio-acoustic surveys based on autonomous audio recording are both cost-effective and time-efficient. However, there are many bird species with different patterns of vocalization, and it is a challenging task to deal with [...] Read more.
Bioacoustics plays an important role in the conservation of bird species. Bio-acoustic surveys based on autonomous audio recording are both cost-effective and time-efficient. However, there are many bird species with different patterns of vocalization, and it is a challenging task to deal with them. Previous studies have revealed that many authors focus on the segmentation of bird audio without considering specific patterns of bird vocalization. Based on the existing literature, currently there is no work on the segmentation of monosyllabic and multisyllabic birds, separately. Therefore, this research addresses the aforementioned concern and also proposes a collection of audio features named ‘Perceptual, Descriptive, and Harmonic Features (PDHFs)’ that gives promising results in the classification of bird vocalization. Moreover, the classification results improved when monosyllabic and multisyllabic birds were classified separately. To analyze the performance of PDHFs, different classifiers were used in which Artificial neural network (ANN) outperformed other classifiers and demonstrated an accuracy of 98%. Full article
(This article belongs to the Special Issue Advances in Machine Learning)
►▼ Show Figures

Figure 1

13 pages, 3566 KiB  
Article
Deep Residual Dense Network for Single Image Super-Resolution
by Yogendra Rao Musunuri and Oh-Seol Kwon
Electronics 2021, 10(5), 555; https://doi.org/10.3390/electronics10050555 - 26 Feb 2021
Cited by 15 | Viewed by 2849
Abstract
In this paper, we propose a deep residual dense network (DRDN) for single image super- resolution. Based on human perceptual characteristics, the residual in residual dense block strategy (RRDB) is exploited to implement various depths in network architectures. The [...] Read more.
In this paper, we propose a deep residual dense network (DRDN) for single image super- resolution. Based on human perceptual characteristics, the residual in residual dense block strategy (RRDB) is exploited to implement various depths in network architectures. The proposed model exhibits a simple sequential structure comprising residual and dense blocks with skip connections. It improves the stability and computational complexity of the network, as well as the perceptual quality. We adopt a perceptual metric to learn and assess the quality of the reconstructed images. The proposed model is trained with the Diverse2k dataset, and the performance is evaluated using standard datasets. The experimental results confirm that the proposed model exhibits superior performance, with better reconstruction results and perceptual quality than conventional methods. Full article
(This article belongs to the Special Issue Advances in Machine Learning)
►▼ Show Figures

Figure 1

15 pages, 775 KiB  
Article
A GPU Scheduling Framework to Accelerate Hyper-Parameter Optimization in Deep Learning Clusters
by Jaewon Son, Yonghyuk Yoo, Khu-rai Kim, Youngjae Kim, Kwonyong Lee and Sungyong Park
Electronics 2021, 10(3), 350; https://doi.org/10.3390/electronics10030350 - 2 Feb 2021
Cited by 4 | Viewed by 2664
Abstract
This paper proposes Hermes, a container-based preemptive GPU scheduling framework for accelerating hyper-parameter optimization in deep learning (DL) clusters. Hermes accelerates hyper-parameter optimization by time-sharing between DL jobs and prioritizing jobs with more promising hyper-parameter combinations. Hermes’s scheduling policy is grounded on the [...] Read more.
This paper proposes Hermes, a container-based preemptive GPU scheduling framework for accelerating hyper-parameter optimization in deep learning (DL) clusters. Hermes accelerates hyper-parameter optimization by time-sharing between DL jobs and prioritizing jobs with more promising hyper-parameter combinations. Hermes’s scheduling policy is grounded on the observation that good hyper-parameter combinations converge quickly in the early phases of training. By giving higher priority to fast-converging containers, Hermes’s GPU preemption mechanism can accelerate training. This enables users to find optimal hyper-parameters faster without losing the progress of a container. We have implemented Hermes over Kubernetes and compared its performance against existing scheduling frameworks. Experiments show that Hermes reduces the time for hyper-parameter optimization up to 4.04 times against previously proposed scheduling policies such as FIFO, round-robin (RR), and SLAQ, with minimal time-sharing overhead. Full article
(This article belongs to the Special Issue Advances in Machine Learning)
►▼ Show Figures

Figure 1

13 pages, 1031 KiB  
Article
Differentially Private Actor and Its Eligibility Trace
by Kanghyeon Seo and Jihoon Yang
Electronics 2020, 9(9), 1486; https://doi.org/10.3390/electronics9091486 - 10 Sep 2020
Cited by 4 | Viewed by 2449
Abstract
We present a differentially private actor and its eligibility trace in an actor-critic approach, wherein an actor takes actions directly interacting with an environment; however, the critic estimates only the state values that are obtained through bootstrapping. In other words, the actor reflects [...] Read more.
We present a differentially private actor and its eligibility trace in an actor-critic approach, wherein an actor takes actions directly interacting with an environment; however, the critic estimates only the state values that are obtained through bootstrapping. In other words, the actor reflects the more detailed information about the sequence of taken actions on its parameter than the critic. Moreover, their corresponding eligibility traces have the same properties. Therefore, it is necessary to preserve the privacy of an actor and its eligibility trace while training on private or sensitive data. In this paper, we confirm the applicability of differential privacy methods to the actors updated using the policy gradient algorithm and discuss the advantages of such an approach with regard to differentially private critic learning. In addition, we measured the cosine similarity between the differentially private applied eligibility trace and the non-differentially private eligibility trace to analyze whether their anonymity is appropriately protected in the differentially private actor or the critic. We conducted the experiments considering two synthetic examples imitating real-world problems in medical and autonomous navigation domains, and the results confirmed the feasibility of the proposed method. Full article
(This article belongs to the Special Issue Advances in Machine Learning)
►▼ Show Figures

Figure 1

17 pages, 1184 KiB  
Article
ST-TrafficNet: A Spatial-Temporal Deep Learning Network for Traffic Forecasting
by Huakang Lu, Dongmin Huang, Youyi Song, Dazhi Jiang, Teng Zhou and Jing Qin
Electronics 2020, 9(9), 1474; https://doi.org/10.3390/electronics9091474 - 9 Sep 2020
Cited by 55 | Viewed by 5152
Abstract
This paper presents a spatial-temporal deep learning network, termed ST-TrafficNet, for traffic flow forecasting. Recent deep learning methods highly relate accurate predetermined graph structure for the complex spatial dependencies of traffic flow, and ineffectively harvest high dimensional temporal features of the traffic flow. [...] Read more.
This paper presents a spatial-temporal deep learning network, termed ST-TrafficNet, for traffic flow forecasting. Recent deep learning methods highly relate accurate predetermined graph structure for the complex spatial dependencies of traffic flow, and ineffectively harvest high dimensional temporal features of the traffic flow. In this paper, a novel multi-diffusion convolution block constructed by an attentive diffusion convolution and bidirectional diffusion convolution is proposed, which is capable to extract precise potential spatial dependencies. Moreover, a stacked Long Short-Term Memory (LSTM) block is adopted to capture high-dimensional temporal features. By integrating the two blocks, the ST-TrafficNet can learn the spatial-temporal dependencies of intricate traffic data accurately. The performance of the ST-TrafficNet has been evaluated on two real-world benchmark datasets by comparing it with three commonly-used methods and seven state-of-the-art ones. The Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and Mean Absolute Percentage Error (MAPE) of the proposed method outperform not only the commonly-used methods, but also the state-of-the-art ones in 15 min, 30 min, and 60 min time-steps. Full article
(This article belongs to the Special Issue Advances in Machine Learning)
►▼ Show Figures

Figure 1

14 pages, 2601 KiB  
Article
Selective Feature Anonymization for Privacy-Preserving Image Data Publishing
by Taehoon Kim and Jihoon Yang
Electronics 2020, 9(5), 874; https://doi.org/10.3390/electronics9050874 - 25 May 2020
Cited by 12 | Viewed by 5276
Abstract
There is a strong positive correlation between the development of deep learning and the amount of public data available. Not all data can be released in their raw form because of the risk to the privacy of the related individuals. The main objective [...] Read more.
There is a strong positive correlation between the development of deep learning and the amount of public data available. Not all data can be released in their raw form because of the risk to the privacy of the related individuals. The main objective of privacy-preserving data publication is to anonymize the data while maintaining their utility. In this paper, we propose a privacy-preserving semi-generative adversarial network (PPSGAN) that selectively adds noise to class-independent features of each image to enable the processed image to maintain its original class label. Our experiments on training classifiers with synthetic datasets anonymized with various methods confirm that PPSGAN shows better utility than other conventional methods, including blurring, noise-adding, filtering, and generation using GANs. Full article
(This article belongs to the Special Issue Advances in Machine Learning)
►▼ Show Figures

Figure 1

Show export options expand_more Show export options expand_less
Select all
Export citation of selected articles as:
 
Displaying articles 1-17
Back to TopTop