ver i3, which is based on neural networks. Most of these programs have their roots in the AI research of the 1980s.
Acknosoft, for example, developed a diagnostic and fault-monitoring system called Cassiopée for Boeing's 737 airliners. A maintenance engineer can descr
ibe a problem to Cassiopée via a question-and-answer session and get back a list of the most similar previous cases and their solutions. Clicking on a case brings up a workshop manual with diagrams and part numbers to help fix the problem.
Neuron Data's business-rules-development frameworks provide components for creating and maintaining a rule base and distributing applications via messaging middleware. It has helped organizations such as American Express, Boeing, NASDAQ, and Walt Disney World to centralize all business policies and procedures into a single repository.
The popularity of the Web has added new impetus to classic pattern-recognition algorithms. There's an increasing need to sift intelligently through reams of random data. Autonomy's Agentware i3 uses linear adaptive filters, a neural-net variant, to learn Web site visitors' preferences and present them with news clips and links that match their interests.
Logic Programming Associates' ProWeb Prolog Server package cont
ains a complete Prolog compiler that has the ability to run server-side Prolog programs to perform searches and manipulations on Web data. In one application, the system provides insurance quotes based on a user's answers to an HTML form.
screen_link (37 Kbytes)
ProWeb Prolog Server provides data to workers in the field.
Harvesting the Fruits of AI Research
