A piece in Smart Data Collective, makes some useful points. I think it is no longer a question. I was actively involved the last time we rode the AI hype in the late 1980s, We spent millions on it then, derived more millions in results, but did not get the billions that management was expecting at the time. Once again I am following it closely, and I think there are indications that real value is starting to emerge, like in the IBM Watson effort, that the world may finally be ready for this.
One reason this is more likely to work this time around is the Big Data explosion. More data is available in more forms than ever before. Artificial Intelligence methods are essentially predictive analytical methods. Our own intelligence works this way. We have memory that we search, operate on it with our business logic, which uses pattern recognition, and predict some future state.
We then take that prediction and envelop it some decision we want to influence. Failures can occur at each of these steps. We can lack the data to search, search it poorly, have inadequate logic or not know how to link it to the actual decisions made in the business.
My own observation was that the failures in AI work occurred for two reasons. First, we did not have adequate or complete data to understand the process, and secondly, we did not take sufficient care to blend the results into the decision process.
We are ready to address both failures now, the data is there. The decision process aspect needs to be solved with more care in understanding the connection between prediction and useful decision. We have all the parts to do this now lets put them together.
Am still a believer that this will be useful for any size business.
This post was written as part of the IBM for Midsize Business program, which provides midsize businesses with the tools, expertise and solutions they need to become engines of a smarter planet. More here.
Monday, March 10, 2014
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment