The Power of Deep Learning: Current Research and Future Trends

Abstract

Deep learning, in general, is about multi layered neural networks copying the structure and intellectual procedure of the human mind. Rather than handcrafted features, it permits the procurement of knowledge straightforwardly from information. They relapse mind boggling target works in a nested system, where more complex forms with bigger receptive fields are estimated using less abstract ones. Deep learning additionally makes it conceivable to consider formal domain knowledge and supplant an extensive collection of traditional algorithmic methods with flexible differentiable modules. These all strengthen and empower deep learning and make it adaptable while establishing the connection between the input information and target yield. Research frontiers are presently moving toward the rest of the difficulties. This paper presents a complete overview about deep learning. It illustrates where did deep learning initiate from, what had been accomplished using deep learning, What research areas are currently being investigated via deep learning, and most importantly What are the challenges and open problems of deep learning - as those are the issues, once handled, will lead to achieve the general conscious Artificial Intelligence (AI). The purpose is to empower graduates, practitioners, researchers and fans toward a powerful cooperation in the field of deep learning.