Jennifer Zaino
SemanticWeb.com Contributor

Cognitive psychologist Lars Ludwig has developed an application
dubbed ArtificialMemory as a test bed for his interests in experimenting with
knowledge management techniques and approaches.

Currently in use with a
couple of companies, the application may be the semantic equivalent of
online analytical processing (OLAP), the simplified approach taken by the
business intelligence community to quickly get answers to multi-dimensional
questions. The system aims at creating a very easy way to access and aggregate semantic information — doing what an OWL engineer would call inferencing — so that individual and group knowledge can be set free from documents and the applications that produce them, in order to be reused as needed. ArtificialMemory supports semantic web standards such as RDF and OWL, while taking a very different approach to the idea of inferencing.

“In order to get rid of documents, you have to break the content into pieces and rearrange it in a useful way,” says Ludwig. “For that you need a kind of semantic framework, and that’s the motivation behind ArtificialMemory, to provide this semantic framework in a fashion that lets you really work fast. You don’t have to think about a data model, really. What you do is kind of either instantiate objects that you already know or just describe what you are noting down.”

Users of ArtificialMemory explicate relationships — something is a book or a person, for example — to create a very simple form of classification that lets knowledge easily be reused later, vs. what he says is the more complex semantic web approach of using OWL ontologies to infer knowledge based on what an instance is and to which classification it belongs.

“The problem is that the semantic applications, as far as I understand them and see them, were originally a movement to make the web as we know it machine-readable,” he says. “The problem is that, actually, a fact being extracted from the web or documents–facts and annotations are not really readable to humans. So there is missing an important part here and that is readability for humans.” RDF, XML, and company are good for machines, “but in the end if you step away from automatic fact generation you are running into a problem. You have to get users to annotate facts, statements, and this is something that has not really been thought through by the scientific community as far as I can see.”

Additionally, merging and mapping ontologies is not sufficient to actually aggregating knowledge, because that leads to abstraction. Additionally, origins and context may be hidden, and so potentially suspect, Ludwig argues.

“Research into ontology merging isn’t really about aggregating information, but rather a kind of mixing and integrating and linking knowledge. That’s far from generating general knowledge in terms of aggregation or in terms of seeing qualitative differences. So, actually in the semantic web world they are missing many parts to deliver more and better knowledge to people.” ArtificialMemory, he says, can be thought of as a program that tries to establish a personal knowledge management system using semantic web technologies, and achieves this by applying the technology to a degree that you can easily use it and express yourself in knowledge chunks (rather than documents) that are semantically framed and related. And also it can be thought of as a program that provides cleaner aggregation than you currently find on the semantic web.

ArtificialMemory’s approach makes it a potentially useful tool in building up enterprise knowledge management, he says. There are problems with starting a knowledge management effort based on documents that reflect a common level of understanding, rather than knowledge chunks that reflect individuals’ diverse and diffuse views. Documents are willful means of communication that are limited in a purposeful way-to reflect what the author thinks may be important to a particular audience, even if that means leaving out important bits and pieces of other knowledge he or she has. When the only information to be mined in a knowledge management effort comes from shared documents that reflect a common understanding and viewpoint, opportunities to extract value from all the other knowledge chunks an individual has accumulated, which can be semantically related are lost.

“I would rather see what is different in thought, word and knowledge structure of a person compared to my knowledge structure,” Ludwig says. “This would deliver a totally different picture from what the person could offer willfully, but it could be useful for both participants. A different way of communicating is possible and would be efficient and could lead to a better combination of information and better general knowledge.”

Ludwig says that to actually change your enterprise knowledge management strategy, you need a different technology.

“All the technology you have today in the enterprise focuses on extracting knowledge and kind of dis-integrating it from the people that produce it,” he says. “You have this need for knowledge sharing in a fast-paced economy that really needs innovation, fast innovation. There is a need in the enterprise world for integrated personal knowledge management, because it is for its best and for society’s and the economy’s best. This is the big picture.”

SemTechBiz is Less Than 2 Weeks Away

The Semantic Tech & Business Conference (SemTechBiz) is coming to San Francisco on June 3-7! Join us for case studies, innovative panels, tutorials, and keynotes that will provide you with practical advice, hands-on guidance, and breakthrough approaches to solving business problems with semantic technology. Passes go up $200 at the door. Sign up now and save !