MACHINE LEARNING STRATEGIES FOR CYBER SECURITY
The IT industry often is often accused of popularising buzzwords that promise a panacea to ills that plague the cyber world. While the criticism may not be entirely unjustified, the evolution of technology and breakthroughs do offer benefits and deliver exciting solutions. Machine learning (ML) is probably a buzzword that has caught the fancy of a lot of people. A large number of solutions are advertised as driven by ML. Even when we discount the hype, ML offers novel and exciting possibilities to many challenges in the IT world. Cyber security is one such area with potential for ML implementation.
The objective of this article is to present views on the potential for ML in the context of cyber security, rather than offer an in-depth discussion of ML itself. Introducing the rudiments of ML, however, aids further discussion of the topic. Traditionally, computers have only executed programs, never written programs themselves. Computers have been tools that have enabled humans to gain insights in various disciplines. They have been tools in the hands of humans, not intelligent or autonomous entities that have capabilities to learn by themselves. Computers have not discovered insights into subjects and independently enhanced their own understanding.
With ML, which is a dominant subset of artificial intelligence (AI), this is likely to change. Using ML techniques, computers are able to create insights and hence ‘learn’. The learning is not inherited by nor injected via any external intervention but learned and gleaned from available information and modelling techniques. Learning is not possible without the availability of relevant information.
Oct-Dec 2019 Issue