Malicious software or malware has become a major security concern in this digital age as computer users, corporations, and governments witness an exponential growth in malware attacks. Malware detection solutions adopt Static and Dynamic analysis of malware signatures and behavior patterns that are time-consuming and ineffective in identifying unknown malware. In malware analysis, feature extraction involves static analysis is extracting the features without executing the code, dynamic analysis is deriving features while running the executable, and hybrid analysis is a combined examination of static and dynamic analysis. Deep learning in malware detection builds classification models and identifies more features than conventional machine learning methods in feature extraction to improve accuracy. The advantage of using deep learning in malware detection is automatically extracting high conceptual features from data, and it is well useful in making a malware defense mechanism. Convolutional neural networks, Deep belief networks, Recurrent neural networks, Deep autoencoders, and long short-term memory are the deep learning architectures predominantly applied in malware detection. Advancements in malware detection using deep learning are gray-ware detection, multi-class malware detection, domain-specific model architecture, and many more.