Search

Deep Learning (DL) has experienced considerable reach and success in the number of various application areas in recent years. The modern era of Machine learning has been rapidly developing and extended to most Convolutional fields of practice, as also to some new fields with more number of opportunities. Based on various categories of learning, numerous approaches have been suggested, including supervised, semi-supervised and unsupervised deep learning. The unsupervised deep learning aims to understand transferable image or video representations without manual annotations. Also, unsupervised approaches are needed when patterns that discern abnormal and normal behavior. In this paper, the recent development methods that are emerged in the domain of unsupervised deep learning are discussed. The various developments in the field of Auto Encoder are explained. The Deep learning structure like Convolutional Neural Network (CNN), Recurrent Neural Network (RNN) is considered as a recent method which is in development for improving the accuracy and to perform the classification in an efficient way.

Diabetic Retinopathy is one of the most dangerous disease and should be identified and treated properly at the very early stage. This is usually diagnosed by scanning the interior structure of human eye with modality like optical coherence tomography and color fundus photography. Then the disease is been diagnosed manually by the respective experts which is a time-consuming process. This process should be automated so that the disease can be diagnosed in a faster and efficient way to reduce the human error. More number of researchers have been done based on automating the diagnosing of diabetic retinopathy disease using machine learning and deep learning approach. The most recent and robust techniques are been discussed in this paper.

Photothermal Therapy (PT) is an efficient scanning technique in biological research. However, rapid PT scanning and noise attenuation constitute a critical challenge in a clinical trial. To solve this problem, techniques based on Compression Sensors (CS) have recently been developed that use lower PT calculation costs and excellent accuracy. The majority of image results of prominence techniques depends on rare and inconsistent circumstances. It also takes considerable time to ensure that the photothermal information gathered empirically meets CS requirements. To address these constraints, a Hyperspectral Images Deep Learning (HSIDL-PT) and Maximize-Minimize Algorithm (MMA) method proposed in this paper to efficiently reduce under-represented artifacts in Neural Network (NN). The proposed method is to retrieve high photothermal images which used 5% of the total measuring information. Furthermore, the quantitative evaluation shows that the s HSIDL-PT technique improves performance by 30 percent when compared to the existingZ technique. with mean Structure Similarity Index (SSIM) = 0.974, Peak Signal-to-Noise Ratio (PSNR) = 29.88 dB and Standard Deviation of ± 0.007. Researchers propose MMA with Deep Learning, which would be the outcome of employing the limit for non-convex optimization, ethical and responsible that used the unifying MMA viewpoint; studies demonstrate that the proposed method outperforms the only one depending on linear maximizing.

Despite of several formulations discussed for operation of Directional Over current Relays (DOCRs) in micro grids the coordination of DOCR is yet a great tasked to be solved. A micro grid is a group of generators preferably a renewable energy source and loads interconnected. The modes of micro grid operation, grid connected mode and autonomous mode are to be considered simultaneously to design an optimal protection scheme for the micro grid. Several conflicting objective functions that must be optimized simultaneously during the relay setting process make the coordination problem to be treated as multi-objective. In this work a multi-objective approach yielding better coordination is presented. There are two major objectives as follows; former is the synchronization period between the main and secondary relays, later is the relay operation time. The conventional Multi-Objective Evolutionary Algorithms (MOEAs) may take thousands of function evaluations, which is not practical for real-world problems. Hence surrogates are introduced in those MOEAs to approximate the fitness function. The relay coordination is devised as a constrained multi-objective problem. The Surrogate Assisted Particle Swarm Optimization (SAPSO) Algorithm is applied to accomplish better coordination between the relays. The scheme is validated using standard IEEE systems and the results are compared with the conventional Multi-Objective Particle Swarm Optimization (MOPSO). The optimal values of pickup current and Time Dial Settings are computed for both modes of micro grid operation. The sum of operating time of all relays and the synchronization period between the main and secondary relays are minimized. The Inverted Generation Distance (IGD) is employed as the parameter for computing the performance of the algorithm. Also the convergence characteristic is analyzed. From the results it is proved that SAPSO have good stability than MOPSO and offers better Pareto fronts, convergence and distribution.

With the advent of cloud computing, the affinity between business and technology had increased manifold, allowing users to access IT resources at their convenience through the pay-per-use scheme. With such huge demand surging day to day, the cloud environment must cater to the user requirements flawlessly and also should be rewarding to the providers of cloud service. To maintain its high level of efficiency, there are several challenges that the cloud environment should tackle. One amongst those challenges is the balancing of load. It is one of the primary features of cloud computing that focuses on avoiding the overloading of nodes where there may be idle nodes or nodes with lesser load present at the same juncture. By keeping an effective check on the load several the Quality of Service (QoS) parameters including response time, throughput, resource utilization, energy consumption, cost etc., can be improved, adding to better performance of the entire cloud environment. Even distribution of load among datacenters will contribute to optimal energy consumption and keeps a check on carbon emissions. In this paper we have presented a methodical review on literature pertaining to load balancing strategies that had been proposed in the cloud environment. We had made in-depth analyses of available load balancing techniques and had come up with their advantages, limitations along with the challenges to be addressed by researchers for developing efficient load balancing strategies in the near future. We had also suggested prospective insights about the aspects in load balancing that could be applied in the cloud environment.