Malware threat detection is one of the most challenging tasks in the field of Information Security and the shortage of qualified personnel makes it even harder for people to keep their information secure. Moreover, malware design has evolved continuously, making it even more difficult for people to protect themselves from malware attacks. Thus, it is the need of the hour to improve the existing malware threat detection systems with modern deep learning algorithms. This paper focuses on bringing together a comprehensive study of various deep learning solutions for the detection of malware from its PE file (Portable Executable File) byte streams.