Deep Learning for Classification of Malware System Call Sequences

摘要: The increase in number and variety of malware samples amplifies the need for improvement automatic detection classification variants. Machine learning is a natural choice to cope with this increase, because it addresses discovering underlying patterns large-scale datasets. Nowadays, neural network methodology has been grown state that can surpass limitations previous machine methods, such as Hidden Markov Models Support Vector Machines. As consequence, networks now offer superior accuracy many domains, computer vision or language processing. This comes from possibility constructing higher potentially diverse layers known Deep Learning.