Artificial Intelligence is the concept where the ability of a machine to carry out cognitive tasks for problem-solving is tested.Multiple subfields of AI have emerged in the last few decades including machine learning, deep learning, natural language processing, fuzzy logic and so on. Deep neural networks have achieved considerable success over the past ten years in a variety of fields. However, deep neural networks have significant computing expenses, huge data requirements, and high energy usage. Exploiting deep neural networks in those applications has been extensively researched due to the recent increase in demand for the autonomy of machines in the real world, Since real-time responses are required in these applications, energy and computing efficiency are crucial due to the limited energy source. Spiking neural networks that are physiologically plausible have lately provided a promising solution to these previously unachievable applications.This paper discusses one such subfield of AI, Spiking Neural Network.The paper presents architectural enhancements in SNN to make it more and more sustainable and robust thereby ensuring provision for maximum accuracy in solving variety of problems across the globe. The paper further discusses the various applications of SNN specific to pattern recognition, logic gates implementation and image and sound processing with the reference from numerous review articles.