The term Deep Learning or Deep Neural Network refers to Artificial Neural Networks (ANN) with multi layers. Over the last few decades, it has been considered to be one of the most powerful tools, and has become very popular in the literature as it is able to handle a huge amount of data. The interest in having deeper hidden layers has recently begun to surpass classical methods performance in different fields; especially in pattern recognition. One of the most popular deep neural networks is the Convolutional Neural Network (CNN). It take this name from mathematical linear operation between matrixes called convolution. CNN have multiple layers; including convolutional layer, non-linearity layer, pooling layer and fully-connected layer. The convolutional and fully-connected layers have parameters but pooling and non-linearity layers don't have parameters. The CNN has an excellent performance in machine learning problems. Specially the applications that deal with image data, such as largest image classification data set (Image Net), computer vision, and in natural language processing (NLP) and the results achieved were very amazing. In this paper we will explain and define all the elements and important issues related to CNN, and how these elements work. In addition, we will also state the parameters that effect CNN efficiency. This paper assumes that the readers have adequate knowledge about both machine learning and artificial neural network.

Recently, social touch gesture recognition has been considered an important topic for touch modality, which can lead to highly efficient and realistic human-robot interaction. In this paper, a deep convolutional neural network is selected to implement a social touch recognition system for raw input samples (sensor data) only. The touch gesture recognition is performed using a dataset previously measured with numerous subjects that perform varying social gestures. This dataset is dubbed as the corpus of social touch, where touch was performed on a mannequin arm. A leave-one-subject-out cross-validation method is used to evaluate system performance. The proposed method can recognize gestures in nearly real time after acquiring a minimum number of frames (the average range of frame length was from 0.2% to 4.19% from the original frame lengths) with a classification accuracy of 63.7%. The achieved classification accuracy is competitive in terms of the performance of existing algorithms. Furthermore, the proposed system outperforms other classification algorithms in terms of classification ratio and touch recognition time without data preprocessing for the same dataset.

Medical image analysis is a significant burden for doctors, therefore, it is used to supplement image processing. Many medical images are assumed to be diagnosed as accurately as healthcare experts when the precision of image detection and recognition in an image processing approach matches that of a human being. Artificial Intelligence (AI) based predictive modelling is an important component of many healthcare solutions. This paper develops and implements a neural network-based method for skin cancer prediction to expose the neural network's strength in this field. This method determines which form of deep learning is best for diagnosing diseases with an accuracy exceeds human ability in terms of speed and accuracy, and determines the optimum number of layers and neurons in each layer for both Convolutional Neural network (CNN) and Deep Neural Network (DNN) to obtain the best possible precision. The results of the proposed method showed impressive results, especially for CNN. There is a clear superiority of CNN over DNN. The CNN (which relies on convolution filters) provides great results in extracting features due to the focus on the intended area of the image without the surrounding area region of interest. This led to a remarkable decrease in the number of parameters and the speed of attaining results with higher accuracy. The results indicated that CNN has a high accuracy rate compared with the other existing methods where the accuracy rate of CNN is 98.5%.

There is a huge information stockpile available on the internet. But the available information still throws a stiff challenge to users while selecting the needed information. Such an issue can be solved by applying information filtering for locating the required information through a Recommender System. While using a RS, the users find it easy to curate and collect relevant information out of massive databanks. Though various types of RS are currently available, yet the RS developed by Collaborative Filtering techniques has proven to be the most suitable for many problems. Among the various Recommended Systems available, movie recommendation system is the most widely used one. In this system, the recommendations will be made based on the similarities in the characteristics as exhibited by users / items. The movie recommendation system contains a huge list of user objects and item objects. This paper combines Collaborative Filtering Technique with association rules mining for better compatibility and assurance while delivering better recommendations. Hence, in the process, the produced recommendations can be considered as strong recommendations. The hybridization involving both collaborative filtering and association rules mining can provide strong, high-quality recommendations, even when enough data is unavailable. This article combines various recommendations for creating a movie recommendation system by using common filtering techniques and data mining techniques

Background: Advance communication systems require new techniques for FIR filters with resource efficiency in terms of high performance and low power consumption. Lowcomplexity architectures are required by FIR filters for implementation in field programmable gate Arrays (FPGA). In addition, FIR filters in multistandard wireless communication systems must have low complexity and be reconfigurable. The coefficient multipliers of FIR filters are complicated. Objective: The implementation and application of high tap FIR filters by a partial product reduce this complexity. Thus, this article proposes a novel digital finite impulse response (FIR) filter architecture with FPGA. Method: The proposed technique FIR filter is based on a new architecture method and implemented using the Quartus II design suite manufactured by Altera. Also, the proposed architecture is coded in Verilog HDL and the code developed from the proposed architecture has been simulated using Modelsim. This efficient FIR filter architecture is based on the shift and add method. Efficient circuit techniques are used to further improve power and performance. In addition, the proposed architecture achieves better hardware requirements as multipliers are reduced. A 10-tap FIR filter is implemented on the proposed architecture. Results: The design’s example demonstrates a 25% reduction in resource usage compared to existing reconfigurable architectures with FPGA synthesis. In addition, the speed of the proposed architecture is 37% faster than the best performance of existing methods. Conclusion: The proposed architecture offers low power and improved speed with the lowcomplexity design that gives the best architecture FIR filter for both reconfigurable and fixed applications.

More and more the use of artificial intelligence and data mining techniques established in many fields to solve the problem of classification. This paper consider most new database applications request smart interface to improve effective collaborations in the middle of database and the clients. The most open interfaces for databases must be clever and ready to comprehend characteristic dialect expressions. The overall aim of this study is to look at the importance of using data mining techniques with artificial intelligence in algorithms and applications. We propose a general design for an intelligent database interface. Furthermore, a genuine usage of such a framework which can be connected to any database. One of the fundamental attributes of this interface is space, freedom, which implies that this interface can be utilized with any database. Another aspect of this framework is that it is easy setup. The intelligent interface utilizes semantic coordinating procedure to change natural language query to Structured Query Language (SQL) by depending lexicon and set of creation guidelines. The lexicon comprises semantics sets for tables and sections. The query model is executed and the outcomes are introduced to the client. This interface was initially tested utilizing Supplier-Parts database by using JAVA and the result proves the efficiency of the proposed method in intelligent database system.