Traditional machine learning (ML) approach is based on a two-part system usually: feature extraction and selection module plus trainable classifier module. Since the early 1930s, when Nicolas Rashevsky developed the first model of neural network (McCulloch & Pitts, 1943), many of us have been practicing with artificial neural networks (ANNs) derived or not derived from biological neural networks (BNNs) for decades. Some others have started after convolutional neural networks (CNNs) and deep learning (DL) showed their amazing impact on applications. Some others are following the Big data and data analytics mood.