Eight universities have begun a year-long initiative to train IBM Watson for work in cybersecurity. Will the Jeopardy champ soon police the internet?
On Tuesday, IBM announced that Watson, its cognitive computing system (and former Jeopardy champion), will be spending the next year training for a new job—fighting cybercrime.
Watson for Cyber Security is a cloud-based version of IBM's cognitive computing tools that will be the result of a one-year-long research project that is starting in the fall. Students and faculty from eight universities will participate in the research and train Watson to better understand how to detect potential threats.
SEE: Network Security Policy Template (Tech Pro Research)
The following universities will be a part of the process:
- California State Polytechnic University, Pomona
- Pennsylvania State University
- Massachusetts Institute of Technology
- New York University
- University of Maryland, Baltimore County
- University of New Brunswick
- University of Ottawa
- University of Waterloo
Like many other cognitive systems, Watson learns by digesting large amounts of information. Essentially, the students will train Watson "by annotating and feeding the system security reports and data," according to an IBM press release. This includes data from IBM's X-Force research library, which contains more than 100,000 documented vulnerabilities.
Students will be working with the folks at IBM Security to accomplish this task. According to IBM, the company plans on processing as many as 15,000 security documents, such as intelligence reports, per month during the course of the research.
Additionally, the University of Maryland, Baltimore County, announced that it was further collaborating with IBM to create an Accelerated Cognitive Cybersecurity Laboratory (ACCL) in its College of Engineering and Information Technology, where students and faculty will work with IBM researchers on new cybersecurity initiatives.
Another hope of IBM's is that, by training Watson in cybersecurity, it can help alleviate the skills gap present in the security industry. Being that Watson for Cyber Security can work with unstructured data, it can help security researchers and analysts process more data overall as the average organization only uses 8% of unstructured data. This is despite the fact that unstructured data makes up 80% of the data that is used in attacks. Using natural language processing (NLP), Watson will be able to better sift through text in content like blogs or news reports to determine potential threats.
SEE: 5 companies using IBM Watson to power their business (TechRepublic)
Currently, Watson for Cyber Security is designed to reveal emerging threats and potential ways of dealing with them. Although, there is no word yet on whether or not Watson will eventually be able to stop cyberattacks or prevent them on its own.
The 3 big takeaways for TechRepublic readers
- IBM announced that Watson will be trained by eight universities to better understand how to detect potential cybersecurity vulnerabilities. The official tool is called Watson for Cyber Security, and it is a cloud-based version of the cognitive capabilities that Watson is known for.
- Watson for Cyber Security's big value addition is that it can process and analyze unstructured data used in cyber attacks, which could help address the skills gap in the security industry.
- By continuing to move away from the marketing gimmicks of its past and embrace the serious business capabilities of Watson (e.g., healthcare and security), IBM is making a stronger case for cognitive computing in the enterprise.
- IBM Watson: The inside story of how the Jeopardy-winning supercomputer was born, and what it wants to do next (TechRepublic)
- IBM Watson-powered app aims to make hospital visits less daunting for young patients (ZDNet)
- IBM Watson: A shining example of how to take big data to the next level (TechRepublic)
- How good is IBM Watson at foreign languages? One Spanish bank is about to find out (ZDNet)
- IBM Watson machine learns the art of writing a good headline (TechRepublic)