Research on the algorithm of painting image style feature extraction based on intelligent vision
تحقیق در مورد الگوریتم استخراج ویژگی های سبک تصویر بر اساس دید هوشمند-2021
Because the traditional image feature extraction algorithm does not smooth the image, the success rate of feature extraction is low, the average running time and the false positive rate are increased. In view of the above problems, this paper proposes an algorithm of painting image style feature extraction based on intelligent vision. According to the internal structure of the content image and the painting image, the similarity analysis and the smooth transfer of pixels are carried out, and then the painting image is smoothed with the semi-supervised learning method. On this basis, the similarity rule of painting image style is established, and all the style features are quantified, so as to obtain the self- similarity descriptor of painting image style. Then the similarity coefficient between the painting image and other sample images is calculated, and the similarity matrix is constructed, and the intelligent vision technology is used to complete the extraction of the painting image style features. Experimental results show that this algorithm can effectively reduce the average running time and false positive rate of painting image style feature extraction, and also improve the success rate of feature extraction.© 2021 Published by Elsevier B.V.
Keywords: Painting image style | Feature extraction | Smoothing processing | Semi-supervised learning | Similarity rule | Intelligent visual
Machine learning and AI-based approaches for bioactive ligand discovery and GPCR-ligand recognition
یادگیری ماشین و رویکردهای مبتنی بر هوش مصنوعی برای کشف لیگاند زیست فعال و تشخیص GPCR-لیگاند-2020
In the last decade, machine learning and artificial intelligence applications have received a significant boost in performance and attention in both academic research and industry. The success behind most of the recent stateof- the-art methods can be attributed to the latest developments in deep learning. When applied to various scientific domains that are concerned with the processing of non-tabular data, for example, image or text, deep learning has been shown to outperform not only conventional machine learning but also highly specialized tools developed by domain experts. This review aims to summarize AI-based research for GPCR bioactive ligand discovery with a particular focus on the most recent achievements and research trends. To make this article accessible to a broad audience of computational scientists, we provide instructive explanations of the underlying methodology, including overviews of the most commonly used deep learning architectures and feature representations of molecular data. We highlight the latest AI-based research that has led to the successful discovery of GPCR bioactive ligands. However, an equal focus of this review is on the discussion of machine learning-based technology that has been applied to ligand discovery in general and has the potential to pave the way for successful GPCR bioactive ligand discovery in the future. This review concludes with a brief outlook highlighting the recent research trends in deep learning, such as active learning and semi-supervised learning, which have great potential for advancing bioactive ligand discovery.
Keywords: Molecular representations | GPCR ligands | Drug discovery | Deep learning | Machine learning | Graph convolutional neural networks
Wireless control using reinforcement learning for practical web QoE
کنترل بی سیم با استفاده از یادگیری تقویت کننده برای QoE عملی وب-2020
Wireless networks show several challenges not found in wired networks, due to the dynamics of data transmission. Besides, home wireless networks are managed by non-technical people, and providers do not implement full management services because of the difficulties of manually managing thousands of devices. Thus, automatic management mechanisms are desirable. However, such control mechanisms are hard to achieve in practice because we do not always have a model of the process to be controlled, or the behavior of the environment is dynamic. Thus, the control must adapt to changing conditions, and it is necessary to identify the quality of the control executed from the perspective of the user of the network service. This article proposes a control loop for transmission power and channel selection, based on Software Defined Networking and Reinforcement Learning (RL), and capable of improving Web Quality of Experience metrics, thus benefiting the user. We evaluate a prototype in which some Access Points are controlled by a single controller or by independent controllers. The control loop uses the predicted Mean Opinion Score (MOS) as a reward, thus the system needs to classify the web traffic. We proposed a semi-supervised learning method to classify the web sites into three classes (light, average and heavy) that groups pages by their complexity, i.e. number and size of page elements. These classes define the MOS predictor used by the control loop. The proposed web site classifier achieves an average score of 87% ± 1%, classifying 500 unlabeled examples with only fifteen known examples, with a sub-second runtime. Further, the RL control loop achieves higher Mean Opinion Score (up to 167% in our best result) than the baselines. The page load time of clients browsing heavy web sites is improved by up to 6.6x.
Keywords: Wireless network | Software defined network | Reinforcement learning | Q-Learning | Multi-armed bandit | Quality of Experience
Semi-supervised gear fault diagnosis using raw vibration signal based on deep learning
تشخیص خطای دنده نیمه نظارت شده با استفاده از سیگنال لرزش خام بر اساس یادگیری عمیق-2019
In aerospace industry, gears are the most common parts of a mechanical transmission system. Gear pitting faults could cause the transmission system to crash and give rise to safety disaster. It is always a challenging problem to diagnose the gear pitting condition directly through the raw signal of vibration. In this paper, a novel method named augmented deep sparse autoencoder (ADSAE) is proposed. The method can be used to diagnose the gear pitting fault with relatively few raw vibration signal data. This method is mainly based on the theory of pitting fault diagnosis and creatively combines with both data augmentation ideology and the deep sparse autoencoder algorithm for the fault diagnosis of gear wear. The effectiveness of the proposed method is validated by experiments of six types of gear pitting conditions. The results show that the ADSAE method can effectively increase the network generalization ability and robustness with very high accuracy. This method can effectively diagnose different gear pitting conditions and show the obvious trend according to the severity of gear wear faults. The results obtained by the ADSAE method proposed in this paper are compared with those obtained by other common deep learning methods. This paper provides an important insight into the field of gear fault diagnosis based on deep learning and has a potential practical application value.
KEYWORDS : Deep learning | Gear pitting diagnosis | Gear teeth | Raw vibration signal | Semi-supervised learning | Sparse autoencoder
Probabilistic active learning: An online framework for structural health monitoring
یادگیری فعال احتمالی: یک چارچوب آنلاین برای نظارت بر سلامت ساختاری-2019
A novel, probabilistic framework for the classification, investigation and labelling of data is suggested as an online strategy for Structural Health Monitoring (SHM). A critical issue for data-based SHM is a lack of descriptive labels (for measured data), which correspond to the condition of the monitored system. For many applications, these labels are costly and/or impractical to obtain, and as a result, conventional supervised learning is not feasible. This fact forces a dependence on outlier analysis, or one-class classifiers, in practical applications, as a means of damage detection. The model suggested in this work, however, allows for the definition of a multi-class classifier, to aid both damage detection and identification, while using a limited number of the most informative labelled data. The algorithm is applied to three datasets in the online setting; the Z24 bridge data, a machining (acoustic emission) dataset, and measurements from ground vibration aircraft tests. In the experiments, active learning is shown to improve the online classification performance for damage detection and classification.
Keywords: Damage detection | Pattern recognition | Semi-supervised learning |Structural health monitoring
A SCiForest based semi-supervised learning method for the seismic interpretation of channel sand-body
یک روش یادگیری نیمه نظارت مبتنی بر SCiForest برای تفسیر لرزه ای کانال شن و ماسه -2019
Machine learning technology has been widely applied in the field of seismic interpretation. In most cases, machine learning assisted seismic interpretation is calibrated or constrained bywells.However, due to the limitation of drilling cost, sometimes there are only a few samples can be obtained from the well-points for a specific layer, which is insufficient to guarantee the generalization ability of supervised learning. In this article, we propose a novelty semi-supervised method by combining the unsupervised isolation forest with split-selection criterion (SCiForest) algorithm and the supervised feature selection process together. The key of the proposed method is to be able to make full use of both the self-contained distribution information of multiple seismic attributes and the calibration information of limited well-points at the same time. To highlight the advantages of the proposed method referred to conventional supervised and unsupervised methods, we take the channel identification practice in the western Bohai Sea as a case study for comparison. Further discussion confirms that the proposed method can improve the visibility of channel effectively by fusing the relevant information in amplitude, frequency, and morphological attributes with limited calibration, which may provide a reliable alternative way for further machine learning assisted seismic interpretation.
Keywords: Machine learning | Semi-supervised learning | Pattern recognition | Isolation forest | Seismic interpretation
Energy consumption modelling using deep learning embedded semi-supervised learning
مدل سازی مصرف انرژی با استفاده از یادگیری عمیق یادگیری نیمه نظارت تعبیه شده-2019
Reduction of energy consumption in the steel industry is a global issue where government is actively taking measures to pursue. A steel plant can manage its energy better if the consumption can be modelled and predicted. The existing methods used for energy consumption modelling rely on the quantity of labelled data. However, if the labelled energy consumption data is deficient, its underlying process of modelling and prediction tends to be difficult. The purpose of this study is to establish an energy value prediction model through a big data-driven approach. Owing to the fact that labelled energy data is often limited and expensive to obtain, while unlabelled data is abundant in the real-world industry, a semi-supervised learning approach, i.e., deep learning embedded semi-supervised learning (DLeSSL), is proposed to tackle the issue. Based on DLeSSL, unlabelled data can be labelled and compensated using a semi-supervised learning approach that has a deep learning technique embedded so to expand the labelled data set. An experimental study using a large amount of furnace energy consumption data shows the merits of the proposed approach. Results derived using the proposed method reveal that deep learning (DLeSSL based) outperforms the deep learning (supervised) and deep learning (label propagation based) when the labelled data is limited. In addition, the effect on performance due to the size of labelled data and unlabelled data is also reported.
Keywords: Energy modelling | Intelligent manufacturing | Deep learning | Semi-supervised learning | Data mining
Semi-Supervised Learning Based Big Data-Driven Anomaly Detection in Mobile Wireless Networks
تشخیص ناهنجاری های رانده شده با داده های نیمه نظارت بر اساس داده ها در شبکه های بی سیم سیار-2018
With rising capacity demand in mobile networks, the infrastructure is also becoming increasingly denser and complex. This results in collection of larger amount of raw data (big data) that is generated at different levels of network architecture and is typically underutilized. To unleash its full value, innovative machine learning algorithms need to be utilized in order to extract valuable insights which can be used for improving the overall network’s performance. Additionally, a major challenge for network operators is to cope up with increasing number of complete (or partial) cell outages and to simultaneously reduce operational expenditure. This paper contributes towards the aforementioned problems by exploiting big data generated from the core network of 4G LTE-A to detect network’s anomalous behavior. We present a semi-supervised statistical-based anomaly detection technique to identify in time: first, unusually low user activity region depicting sleeping cell, which is a special case of cell outage; and second, unusually high user traffic area corresponding to a situation where special action such as additional resource allocation, fault avoidance solution etc. may be needed. Achieved results demonstrate that the proposed method can be used for timely and reliable anomaly detection in current and future cellular networks.
Keywords: 5G; 4G LTE-A; anomaly detec tion; call detail record; machine learning; big data analytics; network behavior analysis; sleeping cell
A rejection inference technique based on contrastive pessimistic likelihood estimation for P2P lending
یک روش رد استنباط برمبنای تخمین احتمال بدبینی مخالف برای وام دهی P2P-2018
The majority of current credit-scoring models are built solely on accepted samples and thus cause sample bias. Sample bias is particularly severe in the peer-to-peer (P2P) lending domain due to its comparatively high rejection rate. Reject inference solves sample bias by inferring the possible outcomes of rejected samples and incorporating them into credit score modeling. This study addresses the problem of reject inference in a specific P2P lending domain from the perspective of semi-supervised learning. A novel reject inference method (CPLE-LightGBM) is proposed by combining the contrastive pessimistic likelihood estimation framework and an advanced gradient boosting decision tree classifier (LightGBM). The performance of the proposed CPLE-LightGBM method is validated on multiple datasets, and results demonstrate the efficiency of our proposal. Analysis of the influence of rejection rate on predictive accuracy reveals the usefulness of sampling in rejected datasets.
keywords: Big data applications |Contrastive pessimistic likelihood |Credit scoring |Data analytics |Gradient boosting decision tree estimation |Machine learning |P2P lending |Reject inference |Semi-supervised learning
From big data to knowledge: A spatio temporal approach to malware detection
از داده های بزرگ به دانش: یک رویکرد زمان فضایی به تشخیص نرم افزارهای مخرب-2018
The deployment of endpoint protection has been gradually migrated from individual clients to remote cloud servers, which is termed as cloud based security service. The new para digm of security defense produces a large amount of data and log files, and motivates data driven techniques for detecting malicious software. This paper conducts an empirical study on the log of a real cloud based security service to characterize the occurrence of execut able files in end hosts, which concerns 124,782 benign and 113,305 malicious executable files occurred in 165,549,417 end hosts. The end hosts and the timestamps that an execut able file occurs in provide insights into the distribution of software in wild from spatial and temporal perspectives, respectively. Meanwhile, we investigate the strategies behind the char acterizations, and observe the preferential attachment process and the periodicity of file occurrence in end hosts. The observed different occurrence patterns of benign and mali cious files in end hosts inspire us a new scalable approach to malware detection. We learn from the characterizations that, the associated files shared more spatial and temporal in formation in common are more likely to be same in their labels, either benign or malicious. Thus, we devise a graph based semi-supervised learning algorithm for real-time malware detection by taking into account the spatio-temporal information of the distribution of ex ecutable files. Experimental results demonstrate that our approach increases the performance on malware detection by 14.7% over previous techniques on average.
Keywords: Malware detection ، Data-driven security analysis ، File co-occurrence ، Graph based semi-supervised ، learning ، Content-agnostic