Review of Deep Learning Methods

Deep Learning falls within the realm of artificial intelligence as a subset of Machine Learning. It plays a crucial role in our everyday lives. The field of Deep learning has expanded significantly in recent years and finds application in wide array of contexts. Deep learning (DL) has made its impact in many areas predictive forecasting, voice generation and recognition, audio, video, image generation and recognition, multimedia, computer vision and language processing etc. Deep Learning primarily employs artificial neural network (ANN) to identify patterns and acquire decision making capabilities from them. These neural networks are designed to emulate the functionality of human brain, which possesses a deep architecture. Deep Learning excels in its capacity to learn from extensive datasets. Deep Learning employs Deep Learning Neural Networks (DNN) comprising multiple layers of units with optimized algorithms and architectures. This paper focuses on many generally used deep learning optimization methods in order to enhance accuracy. Additionally, deep convolution networks (CNN), Artificial Neural Networks, Supervised or Discriminative Learning, Feed-forward Neural Networks, Recurrent Neural Networks (RNN), Backpropagation, convolutional neural networks, Autoencoders, LSTM, and generative advertisement networks (GAN) are encompassed in the review.