04.29.10
Are Computers Smart?
[Cartoon]
Our computers control just about everything, but are they really smart? Do computers only do what programmers have taught them to do or are they capable of connecting the dots?
We have all been to animal shows where a trainer will have the animals perform astounding tricks. How many movies have been made of a dog that seems to lead an intelligent life helping out humans? How many futuristic movies have been made of robots becoming so smart they decide they do not need man any longer?
Have we turned the corner? Are computers smart?
From a business perspective, today we may be pushing it a bit to use the word “smart” and “computer” in the same breath. In many cases, computers have been instilled with sufficient understanding to perform outstanding tricks just like well-trained animals. Consider, for example, the robotic rovers on Mars. They are so far away that transmission time between the Earth and Mars takes longer that the time needed to support remote control. They must have the ability to think for themselves. Does this make them smart?
The better word to use for the ability of computers to make judgments without human intervention is “knowledge-based” system. The concepts used in knowledge-based systems have been around for decades. It has a basic premise: the presence of two or more conditions has a probability of leading to another condition. By having a large knowledge-base of these rules, a rules engine can be used to navigate to a recommended outcome.
For example, if the Mars robot sees a rock in its path and the rock is six centimeters high, then, going forward, the robot has a 70% chance of getting stuck. If the robot turns to the right by a 45-degree angle where there is a rock only 2 centimeters high, the robot has only a 2% chance of getting stuck. These types of rules can exist in the hundreds of thousands and be very quickly processed by a rules engine to determine the next action.
There continues to be an enormous effort to study and better understand how the brain processes information. “Neural networks” is one of those areas of study where the building of knowledge is a learning process that captures knowledge and can act upon it. There are some practical applications, but this is limited in the business world.
There is tremendous business opportunity in applying rules-based systems. The rules are provided by the business units that control the business system. The environment to allow the business unit to maintain the rules and the rules-engine are all that IT needs to provide. This puts the “smarts” of the computing system in the hands of the business unit. With this approach, computers can get “smart” by putting the business units in control rather than the programmers in IT. To be “smart” you must Control Information Technology Yourself.
Closing the Business / IT gap
