The Art of Knowledge

 

Technological Issues Towards Knowledge-powered Organizations
By: George A. Vouros

Right away the name of the paper grabs me. But as I delved into it it just seemed to be a lot more of the same old terminology thrown around. A lot of theories about theories that sound like they are being copied right out of someone else's book. The ideas of technology and education, or KM, obviously go hand and hand. Are there uphill battles that need to be addressed? Absolutely. Is looking towards the past and quoting what you have been reading for the past year going to get us there? No. There's no exam besides results after all the research has been done.

Some excerpts:

Towards this aim, people must work at the "knowledge level": get the right information at the right time and at the right form to perform a task in real time, get connected with colleagues that may provide solutions or hints towards solving problems, form groups of people with different areas of expertise and/or different competencies to achieve a shared goal, be equipped with the necessary applications and data to fulfill their tasks and form decisions in real time. More than that, people should be able to provide feedback and share their knowledge, which must be actively and constantly captured, stored, and organized in the background, so as to be exploited in tasks performance and be disseminated to interested colleagues.

That's a decent overview.

O'Leary (1998) and others point that the knowledge that an organization stores in its repository may be categorized to be knowledge about proposals that the organization made in the past (proposal knowledge repository), knowledge concerning news about organizations' specific topics (news knowledge repository), knowledge about the best way of doing things within the organization (best practices knowledge repository), and knowledge about peoples competencies within the organization (experts knowledge repository).

The past is dwelled on entirely too much. What is the use of wanting to define a new way of looking at business and the way people do business if all of our examples come from and are rooted in past experiences. Simply automating the process by which you were once successful and hoping that it will save you money is one of the reasons eLearning is currently failing. Companies are dropping KM initiatives. They drop them because the processes and knowledge acquisition simply goes from a filing cabinet to an XML schema. You can gussy it up however you want, but that's not worth millions of dollars to a company. If you don't change the inherent principles of how the data is retrieved and then received you have changed very little. Simply taking your old folders of competitor information and making them available via a new medium will do little to change the way people think and learn, it will simply make things available without walking across the room.

These may be bases exploited by expert systems, formal representations of argumentation structures that record group decision processes, formal representations of business processes, or database schemas. Ontologies (Gruber, 1994; Guarino, 1996) are expected to play an important role here and work towards this aim has already been done: representation ontologies make explicit the commitments for structuring knowledge bases, and therefore a semantics-preservation mapping between different knowledge-base constructs is needed for integration (which in most of the cases is not obvious). Domain ontologies make explicit the conceptualization of a specific domain and can be used for mapping domain concepts from one knowledge base to the other. This is also not trivial, given that different ontologies may provide different conceptualizations of a domain at different levels of granularity, with different organizing principles and different implementation languages

That's pretty good stuff. That reminds of what Gee wrote about in the cross referencing of his domains. I thnk he called it domains or some other word. But it all has to do with knowledge mapping. The beauty of things like Enterprise Architectue is that it is inherent knowledge mapping, or at least it could be. If the main premise is the liquid transference of information along channels that are predetermined to not only the user but the process, then why can't knowledge mapping and EA interact on a more grand stage. Ahh I know, because knowledge mapping is inherently flawed by human interaction with a system. And there ain't no way they are taking any risks with EA.

The focus is on what we think are the key technologies for implementing an organizational memory: collaborative applications with transparent and collaborative interfaces; information integration and ontologies; knowledge representation of business processes and ontologies; and knowledge acquisition, data mining and discovery…. The challenge is great. Not only for designing and developing the key technologies, but also for devising the interplay between the technologies and for supporting the collaboration between the applications themselves, in the context of an active organizational memory system. A paradigm for such interplay of technologies and systems' collaboration is the development of collaborative tools for ontology construction. Such tools involve a number of agents that collaborate among themselves towards the construction of a commonly agreed ontology.

What technologies could we use now, that could be implemented? We obviously have XML and it's wonders. But are flash, and video conferencing, and whatever that macromedia thing is, really the answer? They don't feel like the answer. Just because we have TiVo doesn't mean the stuff on television or the way we ingest it is that much better.

Large repositories of information are great, having a system that involves the user making changes is great (though time consuming and destined to fail due to actual user interaction), the idea of ontologies that overlap and break down the barriers of classification and personalization of information is great. However, nobody has shown me anything on this. No system, other than the roots of autonomous and grid computing have shown me a glimpse of anything more than a marketing strategy in programming/EA clothes.

These appear to be the guidelines slated for the achievement

• the development of frameworks and programming methodologies for the development of collaborative agents;
  
• the development of ontologies in conjunction with linguistic resources, which will provide the conceptual system and vocabulary for information retrieval and information presentation in multilingual settings;
  
• the development of standards for integrating heterogeneous information sources and for sharing information between them;
  
• the synthesis and presentation of information, so it can be delivered and be utilized by different device types; and
  
• the development of powerful data mining and knowledge discovery tools for dynamically acquiring users' preferences and interests with the minimum user feedback.

That's not bad either. Unfortunately, still no wooden door to knock on.

Towards the end of the article it did a pretty good job in trying to identify ways in which systems should be drawn to attain, retain, categorize, and make efficient information. It did a nice job of laying a foundation. Oddly enough this is the 50th foundation paper I have read on the topic.



Link posted by JVMM : 3:17 PM