A couple weeks ago, just as we were busy running our European Identity & Cloud Conference, we’ve got news from IBM announcing the company’s foray into the area of Cognitive Security. And, although I’m yet to see their solution in action (closed beta starts this summer), I have to admit I rarely feel so excited about news from IT industry.
First of all, a quick reminder: the term “cognitive computing” broadly describes technologies based on machine learning and natural language processing that mimic the functions of human brains. Such systems are able to analyze vast amounts of unstructured data usually inaccessible to traditional computing platforms and not just search for answers, but create hypotheses, perform reasoning and support human decision making. This is really the closest we have come to Artificial Intelligence as seen in science fiction movies.
Although the exact definition of the term still causes much debate among scientists and marketing specialists around the world, cognitive computing solutions in the form of specialized hardware and software platforms have existed for quite some time, and the exponential growth of cloud computing has been a big boost for their further development. In fact, IBM has always been one of the leading players in this field with their Watson platform for natural language processing and machine learning.
IBM Watson was initially conceived in 2005 as a challenge to beat human players in the game of Jeopardy, and its eventual victory in a 2011 match is probably its best publicized achievement, but the platform has been used for a number of more practical applications for years, including business analytics, healthcare, legal and government services. The company continues to build an entire ecosystem around the platform, partnering with numerous companies to develop new solutions that depend on unstructured data analysis, understanding natural language and complex reasoning.
In the hindsight, the decision to utilize Watson’s cognitive capabilities for cyber security application seems completely reasonable. After all, with their QRadar Security Intelligence Platform, IBM is also one of the biggest players in this market, and expanding its scope to incorporate huge amounts of unstructured security intelligence makes a lot of sense. By tapping into various sources like analyst publications, conference presentations, forensic reports, blogs and so on, cognitive technology will provide security analysts with new powerful tools to support and augment their decision making. Providing access to the collective knowledge from tens of thousands sources constantly adapted and updated with the newest security intelligence, Watson for Cyber Security is supposed to solve the biggest problem IT security industry is currently facing – a dramatic lack of skilled workforce to cope with the ever growing number of security events.
Naturally, the primary source of knowledge for Watson is IBM’s own X-Force research library. However, the company is now teaming with multiple universities to expand the amount of collected security intelligence to feed into the specialized Watson instance running in the cloud. The ultimate goal is to unlock the estimated 80% of all security intelligence data, which is currently available only in an unstructured form.
It should be clear, of course, that this training process is still work in progress and by definition it will never end. There are also some issues to be solved, such as obvious concerns about privacy and data protection. Finally, it’s still not clear whether this new area of application will generate any substantial revenue for the company. But I’m very much looking forward to seeing Watson for Cyber Security in action!
By the way, I was somewhat disappointed to find out that Watson wasn’t actually named after Sherlock Holmes’ famous friend and assistant, but in fact after IBM’s first CEO Thomas Watson. Still, the parallels with “The Adventure of the Empty House” are too obvious to ignore :)