![]() |
|
| |
Every business has one or more individuals whose expertise is in short supply -- someone they would like to clone. Until genetics provide that capability, the best we can do is embody that person's expertise into an expert system that can be incorporated into software applications or served up on the Internet or an intranet. The problem is that it's just not that easy to build an expert system. The knowledge differs from one industry to the next, and the reasoning process differs from one company (or individual) to the next. And, the experts with the knowledge often do not have ability to encode it in a software program. KnowledgeWright solves these problems by providing a flexible framework for representing knowledge and a sophisticated toolset for implementing reasoning rules to draw conclusions from that knowledge. At Amzi!, we can use that framework and toolset to clone a human expert for you. ExamplesCole Davis (coledavis@hotmail.com) is an organizational psychologist building a jobs advisor based on a combination of psychological personality trait profiling and conventional constraints. He has a way of representing jobs and their attributes that is the base knowledge of his system. There is a specific process he uses to apply that knowledge. His system is a component of a larger Visual Basic application. Ken Burres is a Doctor and top endurance athlete. His company, FitCentric (www.fitcentric.com), offers a coaching tool for serious athletes. They want to add an expert advisor that makes workout recommendations based on a large number of physiological factors. The purpose is to help an athlete get to his or hers optimal performance. The knowledge for each factor is stored in tables, and draws on a database of athlete information. The tables are combined to make a composite recommendation. The advisor initially will be part of a Visual Basic application and might make it to the web in the future. Amzi! has built custom KnowledgeWright jigs for both of them. What is a KnowledgeWright Jig?A jig is a pattern, like one you would follow with a jigsaw. For KnowledgeWright a jig is the pattern of the knowledgebase and consists of two parts:
Collectively these two items make up a KnowledgeWright Jig. The framework consists of objects, each of which has a set of properties. Knowledge ObjectsKnowledge is represented by objects. And each object has a set of properties. You can think of each object as a 3x5 card where each line has an attribute and one or more values for that attribute. For example, the job advisor needs to represent each possible job. A job consists of propeties such as the personality traits best suited for that job, the educational requirements to perform the job, the life-style attributes specific to that job, and so on. For the work out advisor, the primary object is a physiological factor, for example heart rate, breathing rates, general health, endurance measures and so on. For each factor there is a table that specifies the conditions under which the factor applies and rules for the weight of this factor relative to other factors. The KnowledgeWright Workshop provides an intuitive, easy to use interface for specifying the objects that make up a knowledgebase. Many of the Workshop's functions are driven by the Jig. So to add a new object or a new property, all we have to do is a simple change to the Jig and the Workshop instantly supports it. Reasoning EngineThe second part of the expert system, the reasoning engine is custom tailored for each application. The reasoning engine "reasons over" the objects in the knowledgebase to reach one or more conclusions. The job advisor works to find all the jobs suited for a person, where suited means that first the job matches their personality profile, and second it meets the constraints (education, life-style etc.). The work out advisor combines all the factor tables according to the weights and conditions into an aggregate recommendation. In addition, the system automatically gathers the bits of information it needs, such as current data from the athlete and past performance data from the database. Although these two examples have very different reasoning strategies, both engines share a large set of basic capabilities. These capabilities include the ability to specify conditions or constraints, to specify values for variables, to represent information in tables, to ask questions of the user and to access values from a database. KnowledgeWright has these capabilities and many more. Amzi! can use these capabilities to craft a reasoning strategy that fits your knowledge and the ways you apply that knowledge. Your reasoning engine runs under the KnowledgeWright Workshop. This provides the ability to run your knowledgebase a step-at-a-time (with a comprehensive log) so you can tune your expert system to run exactly the way you want. Run Your Expert System AnywhereSome expert systems are deployed as part of a larger application -- others are put on the web. Many expert systems come both ways. KnowledgeWright has a ready-to-run web interface so that any knowledgebase can instantly be made available on the web. And it has programming interfaces so your expert systems can be included in virtually any application including Java, Visual Basic, C++ and Delphi. Put Your Name on ItYou can repackage and redistribute KnowledgeWright as part of your software applications or services. In addition to building a custom jig, we can create your own Workshop (with documentation and samples) so that your users or customers can build and use expert systems of your own design. Tell Us What You NeedTo get your own custom jig, tells us about the structure of your knowledge and reasoning processes you use on that knowledge. We'll tell you want it takes to build it. Because of our Jig architecture, most of the software is already written, so we can build your jig quickly and cost efficiently. And with our 10 years in business, we'll be there to enhance and maintain your jig for years to come.
Copyright ©1995-2003 Amzi! inc. All Rights Reserved. Privacy Policy. Amzi!, KnowledgeWright and WebLS are registered trademarks and Adventure in Prolog, Logic Server, Logic Explorer, Adventure in Prolog and Subscription Plus are trademarks of Amzi! inc. |