Data processing: artificial intelligence – Knowledge processing system
Reexamination Certificate
1999-07-27
2003-03-04
Voeltz, Emanuel Todd (Department: 2121)
Data processing: artificial intelligence
Knowledge processing system
C706S046000, C706S047000
Reexamination Certificate
active
06529889
ABSTRACT:
FIELD OF THE INVENTION
This invention refers to the capture, organization, storage and presentation of expert knowledge and the facilitation of the interaction with that knowledge by other professionals. This invention also produces end-products relating to the professional's activities.
BACKGROUND OF THE INVENTION
An expert is considered to be someone who has extensive knowledge on a given topic. Traditionally, if an expert wanted to record his knowledge into a usable format, he would write a book or article, put thoughts into a diary or file, etc. This “published” knowledge would then be available for use by others who could learn from it and utilize it.
In the 1970s and 1980s, the continuing development of computing technique and power resulted in a software category of “expert systems”. The purpose behind expert systems is to take the expert's knowledge of a situation, event, circumstance, etc. and translate that into a software application, usable by other professionals who are working on a similar problem. In using the expert system, the professional follows through the series of steps, as designated by the software, to both “correctly” apply an expert process against the event under consideration and to reach an appropriate conclusion to that event.
The key to the expert system is that the steps that the software presents to the user/professional are based on a series of rules whereby the conduct or answers within one step determines what the next step should be. Once these rules are established within the software application, the user/professional is required to follow the steps as they are presented to him in order for the software (and therefore the user/professional) to function correctly. The user/professional found that differences in approach, timing, unaccounted factors, etc. led to so many exceptions to the stepwise rules of the expert system, that the system became essentially ineffective.
As well, the rules of presentation for the process often closely reflect the personal strategic approach to the situation of the authoring expert. This means it does not necessarily meet the strategic needs and expectations of the professionals who uses the expert system.
Furthermore, any changes to the expert system to reflect alternative models, unaccounted circumstances, and just general evolution of knowledge about the event, are very labor intensive and expensive to perform, leaving the expert system essentially static once it has been implemented. Over time, the expert system is also increasingly limited in utility. The user/professional has determined that expert systems lack important flexibility and are lacking from a conceptual standpoint. As a result, expert systems have failed.
Having recognized the shortcomings of expert systems, the software industry introduced Decision Support Software (DSS). In DSS, the user is asked to gather facts and supply that information to the software application. The DSS then uses that information to determine the likelihood that the professional can expect a particular outcome or may be dealing with a certain type of event. However, the wizened professional knows the factors and outcomes before they begin inputting information into the DSS. That makes the use of the DSS a trite exercise. Then when the DSS is needed for support on more complex issues, the professional finds the DSS fails to account for many of the possible factors that contribute to that particular situation. It also comes up short in recognizing the many permutations that the situation entails.
The DSS strategy, much like the expert system, is expensive and costly to maintain and upgrade. DSS also does not provide the required necessary flexibility because it too is rule bound in determining likely answers to a problem or outcomes of events. So it, too, is a conceptual failure except in cases where a DSS is required because expertise is rare and/or the financial backing is strong allowing the DSS knowledge base maintenance and upgrade cycles to continue.
Stepping into the gap left by expert systems and decision support software is specialty report writing software products that attempt to alleviate the administrative and bureaucratic burden faced by professionals. The strategy behind these software applications is that the professional answers a series of questions after which the software generates a narrative report. However, these software applications have drawn too closely on expert systems approach and have therefore inherited their technological shortcomings, i.e., predetermined step-wise use and lack of flexibility in content coverage.
Specialty report writers introduce a new problem. The narrative generation capability is usually built around standard or stock paragraphs or sentences where the answers to the questions determine the text to pull in or merge into the final document. The strategy here is a sophisticated hybrid of word-processing cut and paste and mail merge technologies. These strategies work best when the resulting documents are standard forms for an industry. Legal, financial, and insurance disciplines have benefited most from this software category.
To make this strategy work, programmers must build an infrastructure for document generation including the construction of databases, user interfaces including forms which display all the questions, and the building of the merge functions which drive the document generation. Much like expert systems and decision support, development, maintenance, and upgrade cycles are very expensive. The resulting software is, once again, relatively static and limits the creative approach of the user/professional. And they produce rigid, stratified final documents that limit the number of disciplines to which this kind of solution finds applicability.
With no other solution in sight, most professionals are left to their word-processors in order to produce their necessary communications. Unfortunately, communication skills vary across any given professional group. Further complicating the issue is that management must interpret the content of a report because different writers/professionals use a multiplicity of terms and different approaches when writing a report—even if the facts are the same. Comparison of reports across different situations is therefore extremely laborious. As well, each professional, depending on training, experience, and specialty, may focus on different aspects of a situation leaving gaps in coverage. In fact, in many cases the time and effort necessary to fully document a situation or event is an increasingly unrealistic and expensive burden. It is therefore difficult for management or other recipients, in reviewing a compiled report, to potentially catch the full understanding of the situation.
This problem has therefore become one of managing information and knowledge. The first attempts at managing knowledge can be referred to as disaggregation strategies. By breaking the professional process into its components, in essence the facts or information the professional was gathering, and perhaps the conclusions they were drawing, organizations have turned the professional process into information that can be readily accessed and shared. Put more simply, organizations develop checklists for their professionals to complete.
This strategy has serious downsides. First, to computerize the checklists still requires the expense of programmers to build the data structures and the interfaces. Again, changes are time consuming and costly.
Furthermore, the report writing side of the process is usually eliminated because of its inherent complexity. This has two repercussions: 1) the professional has to complete the checklist but still has to write a report. There is no time saving in this process, only a doubling of the work; 2) or only checklist completion is required for task completion. This strategy ultimately reduces the scope and amount of information gathered and produces less knowledge about events as opposed to more.
This strategy for managing knowledge therefore falls short.
Mor
Bromberg David
Roberts Stephen
Acappella Software, Inc.
Holmes Michael B.
Lackenbach & Siegel LLP
Todd Voeltz Emanuel
LandOfFree
System and method of knowledge architecture does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with System and method of knowledge architecture, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and System and method of knowledge architecture will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3029251