In the last few years, the deep learning (DL) computing paradigm has been deemed the Gold Standard in the machine learning (ML) community. Moreover, it has gradually become the most widely used computational approach in the field of ML, thus achieving outstanding results on several complex cognitive tasks, matching or even beating those provided by human performance. One of the benefits of DL is the ability to learn massive amounts of data. The DL field has grown fast in the last few years and it has been extensively used to successfully address a wide range of traditional applications. More importantly, DL has outperformed well-known ML techniques in many domains, e.g., cybersecurity, natural language processing, bioinformatics, robotics and control, and medical information processing, among many others. Despite it has been contributed several works reviewing the State-of-the-Art on DL, all of them only tackled one aspect of the DL, which leads to an overall lack of knowledge about it. Therefore, in this contribution, we propose using a more holistic approach in order to provide a more suitable starting point from which to develop a full understanding of DL. Specifically, this review attempts to provide a more comprehensive survey of the most important aspects of DL and including those enhancements recently added to the field. In particular, this paper outlines the importance of DL, presents the types of DL techniques and networks. It then presents convolutional neural networks (CNNs) which the most utilized DL network type and describes the development of CNNs architectures together with their main features, e.g., starting with the AlexNet network and closing with the High-Resolution network (HR.Net). Finally, we further present the challenges and suggested solutions to help researchers understand the existing research gaps. It is followed by a list of the major DL applications. Computational tools including FPGA, GPU, and CPU are summarized along with a description of their influence on DL. The paper ends with the evolution matrix, benchmark datasets, and summary and conclusion.

/pdf/review-of-deep-learning-concepts-cnn-architectures-hpbk62qzc2.pdf

Review of deep learning: concepts, CNN architectures, challenges, applications, future directions

/pdf/deep-learning-in-medical-image-analysis-ilst8as5f9.pdf

Deep Learning in Medical Image Analysis

Performance of deep learning vs machine learning in plant leaf disease detection

A deep learning approach to detect Covid-19 coronavirus with X-Ray images.

Hybrid deep learning for detecting lung diseases from X-ray images

MIMO denotes multiple input multiple output system. Massive MIMO technology can be considered one of the crucial technologies for the future generation of wireless communication systems such as 4G, 5G etc.. The execution and fulfillment of massive MIMO systems has been analyzed in this paper considering M number of antenna users. The advantage is that M streams can be multiplexed per user. The necessary mathematical equations of spectral efficiency in both cases (uplink and downlink) have been derived for any M. The expressions are attainable with the help of channel estimation and per-user-basis MMSE-SIC detectors. MMSE SIC means Minimum Mean square error with successive interference cancellation. Large-system approximations of the spectral efficiency have been obtained. From this illustrative discussion it can be declared that MMSE-SIC has similar asymptotic spectral efficiency (SE) as linear MMSE detectors, which indicates that the SE expanded from users having multiple antennas can be obtained using linear detectors. The power scaling theorems have been generalized in order to handle the M users. It has also been focused that the user can diminish the multiplication of the pilot power and payload power which is the reciprocal of the number of base station antennas.

Uplink and Downlink Spectral Efficiency Estimation for Multi Antenna MIMO User

Visualization and prediction of energy consumption in smart homes

Credit card is a critical element in digital banking. As modern technology is evolving day by day, credit card fraud is increasing significantly. Many financial companies lose billions of dollars annually because of credit card fraud. Criminals constantly try to discover various rules to require illegal actions. As a result, fraud detection methods are crucial for non-banking and banking financial institutions for minimizing the losses. This paper mainly discusses the performance of random forest (RF), AdaBoost, CatBoost classifiers for classification of credit card fraudulent activities. In this paper, the most important features are also selected by applying these classifiers. Results show that among the classifiers considered, RF and CatBoost have the best performance obtaining 99.92% accuracy in detecting credit card frauds.

Effective Machine Learning Approaches for Credit Card Fraud Detection

IoT Driven Healthcare Monitoring System

This paper mainly studies about the prediction of energy consumed by appliances that have been generally used in the house for low energy. Weather records from an adjacent station were collected in order to develop the forecast of the dataset. Dataset used in this paper contains measurements of temperature besides humidity measuring device from a wireless system, climate from an adjacent airport station, and recorded energy consumption of illumination features. Temperature and humidity try to increase the prediction accuracy. This paper discusses some classifiers to predict the humidity and classify the data according to the attributes. Five classification representations were trained with reiterated cross-validation and calculated in the dataset: (a) KNN, (b) support vector machine, (c) SGD, (d) random forest, and (e) neural network. In this paper, sieve diagram has been introduced for comparing the classifiers with one another which is a crucial contribution to our paper. Python base orange data mining tools have been used in this paper. In simulation result, box plot represents standard deviation (SD) and median and epitomizes the values between the first quartile and the third quartile. The Sieve diagram describes the difference between observed and expected frequency whrere it looks like as the density of covering whether the deviation or divergence from independence is negative (red) or positive (blue). Scatter plot shows regression line and regression value which is compared with each other.

https://link.springer.com/content/pdf/10.1007%2F978-981-13-9920-6_25.pdf

Subrato Bharati

Papers

Uplink and Downlink Spectral Efficiency Estimation for Multi Antenna MIMO User

Visualization and prediction of energy consumption in smart homes

Effective Machine Learning Approaches for Credit Card Fraud Detection

IoT Driven Healthcare Monitoring System

Prediction of Energy Consumed by Home Appliances with the Visualization of Plot Analysis Applying Different Classification Algorithm