Watson Applies Machine Learning to Cybersecurity

CYBR650 - Week 9

By learning you will teach, by teaching you will learn. – Latin Proverb

I'm working on artificial intelligence. Actually, natural language understanding, which is to get computers to understand the meaning of documents. -- Ray Kurzweil, computer scientist and futurist

IBM announced earlier this week that they were beginning a new year-long cybersecurity initiative with eight North American Universities.  These universities all have strong cybersecurity programs.  The university students will be working to train and then analyze security data for trends using IBM’s Watson technology platform. 

http://www.wallpaperhd.pk/machine-learning-smart-brain-wallpaper/

IBM’s Watson is a cognitive computing system and former Jeopardy champion.  Watson uses machine learning to extrapolate patterns from large data sets.  The process is to allow the computer to learn from unstructured data, then to make predictions based on its programming, then repeat the process.  This produces reliable, repeatable results.  It’s very similar to artificial intelligence.  Machine learning uses large amounts of unstructured data (in this case, on a specific topic) and makes predictions, tests them against the data and then modifies the algorithm and repeats the “learning” process again.  As new data is added, the system adjusts.  This is one of the methods Google used to program its self-driving car.  The Google car used machine learning to understand how other vehicles behave on the road with 70,000 miles of driving data.  Recording how other drivers and different types of vehicles respond to slow vehicles, obstructions in the road and other situations.  The car learns from others and from its own actions and constantly adjusts the rules that it uses to drive.  The more the car drives in an area, the better it does.

In the case of Watson’s cybersecurity machine learning, the students will train Watson by feeding large amounts of security reports, and other unstructured data.  This includes IBM’s X-Force research library which contains 100,000 documented vulnerabilities.  Watson uses natural language processing to read blog posts, news reports and other information.  The expectation is that Watson for Cybersecurity will reveal emerging threats and how to deal with them.

The project won’t start until fall of this year, so we don’t know if this will revolutionize cybersecurity, but IBM believes this will strengthen the case for Watson’s cognitive computing as a serious business platform.  It may also help with the skills gap of training new cyber professionals.