Computing is changing. Though the changes are gradual, the end result will be to transform computers from essentially dumb devices capable only of crunching numbers with extraordinary speed to devices that combine that skill with intuition and common sense.
The most visible example right now is IBM’s Watson, which famously beat Jeopardy champ (or, human champ) Ken Jennings in 2011. Big Blue has shown during the past couple of years that it sees Watson as far more than a PR triumph. It has commercialized the Watson platform in a number of verticals.
At the highest level, the goal of cognitive computing is pretty intuitive: The idea is to make computers think like humans.
What this precisely means and entails, however, is more complicated and fuzzy. Sue Feldman and Hadley Reynolds discuss the topic in depth at KM World. Their conclusion is that cognitive computers must learn as information changes (adaptive); be interactive; be able to solve problems by asking questions (iterative); “remember” previous interactions (be stateful) and understand context.
In short, they must act like people. Or, at least like some people.
The area appears to be picking up speed. Last week, Cognitive Scale, which is filled with ex-IBM Watson employees, came out of stealth mode. eWeek says that Cognitive Scale has built a new platform based on IBM Watson. It offers applications in four verticals, owns 22 patents, and already has customers and established partnerships with what the story calls “technology titans.” One of these is IBM.
Quotes in the story point to three statistics that illustrate the problems Cognitive Scale aims to solve, according to founder Matt Sanchez: 55 percent of Big Data projects fail, 70 percent to 80 percent of the data “is trapped in silos within and outside company walls with no secure and reliable way to access it,” and 80 percent of it is not readable by machines. Computers that have both the smarts to ferret out that data and the muscle to crunch it will be extraordinarily useful.
TechRepublic goes into some detail about how IBM Watson and other cognitive computer systems can be brought up to speed:
Companies wanting to tap Watson's ability to learn from natural language in documents and questions can plug applications into the seven APIs IBM has made available. These interfaces expose Watson's core ability to provide credible answers to everyday questions, alongside services ranging from machine translation to user modelling, which categorises people based on their email or public posts.
The two key steps are providing IBM Watson with comprehensive information and training it in what seems to be about the same way as one would train an employee.
Carl Weinschenk covers telecom for IT Business Edge. He writes about wireless technology, disaster recovery/business continuity, cellular services, the Internet of Things, machine-to-machine communications and other emerging technologies and platforms. He also covers net neutrality and related regulatory issues. Weinschenk has written about the phone companies, cable operators and related companies for decades and is senior editor of Broadband Technology Report. He can be reached at firstname.lastname@example.org and via twitter at @DailyMusicBrk.