Machine learning

Fig. 1: Can robots learn?

It is universally acknowledged that machines cannot be considered intelligent unless they are able to increase their knowledge and improve their abilities. One way to solve this problem, if only partially, is to provide symbolic machines with inductive and deductive means of  reasoning. Inductive reasoning proceeds from single statements which are expressed through hypothesis and theories to explain the facts provided and may be able to predict new ones. While deductive inference preserves the truth (in the sense of logical correctness), inductive inference does not guarantee it, and so such systems may tend to an excessive generalisation and finally produce mistakes.

For example, an artificial system might learn the concept that "any animal with wings can fly", because it has only encountered examples of flying animals with wings. But there are counter examples: Ostriches have wings but cannot fly.

It is important to note that whilst the process of inductive inference can produce errors (such as the one cited above), deductive inference (applied in automatic proof systems) preserves the "Truth" (i.e., the logic process is sound).

One of the most well known programs capable of learning from examples is ID3, developed by J. Ross Qunlan (between 1979 and 1983), which gave birth to commercial products capable of automatic classification.

Presently, learning programs are used mainly in practical operations in order to meet the need to make use of the wealth of information contained in great data collections accessible through the net, or in company data bases, to reveal a pattern within the data, to extract information and hidden knowledge (data mining).

The Webweavers: Last modified Wed, 09 Mar 2005 11:04:43 GMT