|
USING CONTEXT TO DETERMINE THE BEST SOLUTION
Context - and the way it changes meaning and interpretation - has intrigued Victor Hansen for as long as he can remember. At best, he maintains context can revolutionize information systems. At least, he feels it will generate key new business intelligence applications that dramatically enhance executive decision making. Hansen began thinking about business intelligence as a student at MIT where he studied electrical engineering, natural language and Artificial Intelligence (AI). Today, his pioneering work in this area continues as president and driving force behind MKS, where he designs sophisticated software solutions for business intelligence applications. The notion of context is central to Hansen's contention he can use computers to process large amounts of data more like people. And he's found a quick and efficient way to apply context to common applications with a new software tool called V4. "I've always believed that if we could simply add a layer of context between the users of a very large database itself, then we could revolutionize the way people interact with computers," he said. "We could use sophisticated 'what-if' queries, using context, and actually learn new things that typical OLAP functionality cannot provide. "With context, we can make it easier and more efficient to browse the web, as well as enhance R&D and manufacturing. And we can help executives learn new things about their businesses and competitors that will give them real competitive advantages." Hansen's argument seems to make sense. We, as human beings, are constantly receiving sensory data from the environment. We take the data and interpret it in the context of a very brief, very specific moment in space and time. Everybody interprets differently. Life takes on different meanings and significance, bases on the contextual scenario. What if computers worked this way too? The promise of this type of next-generation, computer functionality - context - is the driving force behind V4. While MKS is currently marketing its contextual tool as an OLAP solution for several very niche-specific applications, Hansen feels very strongly that the following areas are prime targets for "contextual" solutions: REPRESENTATION OF IDEAS -- Complex ideas that have bedeviled AI researchers appear to have representatives in the V4 multidimensional model. For instance, conflicting facts can be represented. Different points of view can be represented ("I know that John thinks Mary is crazy. Is Mary crazy?") The concept of time can also be represented as another contextual dimension in this model. Metadata and data can coexist, permitting the auto-construction of new routines based on the active data model. Words can be given multiple semantic and syntactic meanings based on the current context. RECOMMENDER SYSTEMS -- How do you get information off the web quickly and efficiently? By entering a simple word query on Yahoo or Alta Vista, it's not uncommon to get over 10,000 responses. With a contextual emphasis, queries for information may be viewed from the perspective of personal background, profession, interests, education levels, etc. With a push capability like this, user can tell the Web what context they want to view in, then see only what is appropriate. EXPERT SYSTEMS -- An expert system can mean a lot of different things to a lot of different people. Hansen's general interpretation of an expert is someone or something that figures out a solution or answer within a specific context. It would be an interesting challenge to design an expert system that used an optimization technique to iterate through a problem space represented as context points to determine the "best" solution. At present, Hansen's development team at MKS are at the vanguard of the contextual movement. "There are very few people talking about context or even writing about it in academic journals," Hansen concludes. "The technology is advances, but it may just be the first step in a generation of solutions, tools and products that will have a dramatic impact on information systems and computing for many years to come.
May/June 1997 Software Developer & Publisher |
|
|
MKS,
Inc. • 987 Old Eagle School Road • Wayne, PA
19087-1708 |
"IF
YOU'RE LOOKING FOR FLYING HORSES, YOU'RE NOT GOING TO FIND
TOO MANY AROUND." BEGINS VICTOR HANSEN, PRESIDENT OF MKS
SOFTWARE IN WAYNE, PA. "OF COURSE, HORSES DON'T FLY. BUT,
WHAT IF YOU CHANGE THE CONTEXT OF YOUR QUERY? TO GREEK
MYTHOLOGY, FOR EXAMPLE NOW YOU GET PEGASUS, THE FLYING
HORSE."