K-Now: Here and Now
December 17, 2008
Guest Feature by Dawn Marie Yankeelov, AspectX.com
I have been discussing progress in semantic knowledge structures with Entrepreneur and Researcher Sam Chapman of K-Now who has recently left the University of Sheffield, Department of Computer Science, in the United Kingdom to go full-time into the delivery of semantic technologies in the enterprise. His attendance at the ISWC 2008 has created some momentum to engage new corporations in a discussion on a recently presented paper on “Creating and Using Organisational Semantic Webs in Large Networked Organisations” by Ravish Bhagdev, Ajay Chakravarthy, Sam Chapman, Fabio Ciravegna and Vita Lanfranchi. Knowledge management has shifted as evidenced in his paper. He contends with others that a more localized approach based on a particular perspective of the world in which one operates is far more useful than a centralized company view. All-encompassing ontologies are not the answer, according to Chapman. In the paper, his team indicates:
A challenge for the Semantic Web is to support the change in knowledge management mentioned above, by defining tools and techniques supporting: 1) definition of community-specific views of the world; 2) capture and acquisition of knowledge according to them; 3) integration of captured knowledge with the rest of the organisation’s knowledge; 4) sharing of knowledge across communities.
At K-Now, his team is focused upon supporting large scale organizations to do just this:capturing, managing and storing knowledge and its structures, as well as focusing upon how to reuse and query flexible dynamic knowledge. Repurposing trusted knowledge in K-Now is not based on fixed corporate structures and portal forms, but rather from capturing knowledge in user alterable forms at the point of its generation. Engineering forms, for example, that assist in monitoring aerospace engines during operations worldwide can be easily modified to suit differing local needs. Despite such modifications being enabled this still captures integrated structured knowledge suitable for spotting trends. Making quantitative queries without any pre-agreed central schemas is the objective. This is possible, under K-Now’s approach, due to the use of agreed semantic technology and standards.
The semantic storage K-Now uses (in common with other semantic companies) is based on standardized specifications such as RDF, representing knowledge as a graph of atomic facts. Such a structured representation is machine processable, making knowledge reusable, and easy to exchange, Chapman points out.
As an example, business forms relating to the operation of aerospace engines gives access to information to other unrelated engineers. This helps identify potential improvement opportunities regardless of the source of the knowledge. If a business need for new knowledge occurs or a business process changes then, Chapman contends, it is straightforward to simply design new or alter existing forms. These represent the new knowledge making large organizations far more dynamic than those using static IT solutions, he added.
K-Now’s approach is based on letting users (given access rights) alter knowledge structures themselves. This flexibility allows users to use knowledge locally from their own perspective and in the way they need. K-Forms is also integrated with K-Search which allows querying knowledge in a flexible way using multiple perspectives. Users can query using keywords but also use structured semantically encoded concepts in combination thus guaranteeing more freedom and flexibility to suit all information needs.
K-Now not only focuses upon capture and query of knowledge but also uses semantic technology throughout their applications. Another product, K-Integrate, for example powers knowledge integration bringing together information repositories to align knowledge for reuse and flexible analysis. This technology is proving especially useful for transactional records where the “return-on-investment” question is clearly answered. Looking at corporate procurement records and integrating the knowledge between departmental and sub-organizations it is possible to identify a means bring 5 to 10% in savings to a company’s procurement costs.
K-Now like many semantic companies has a challenge of applying its new technology within the enterprise due to legacy processes and systems. K-Now address this problem by ensuring ease of integration and adoption of its new technology. For legacy data K-Now use K-Extraction (natural language processing and rule based extraction techniques) to automatically align legacy data to ease adoption.
As for open-source solutions, Chapman concurs that we will see more of that in the enterprise—structured resources will become increasingly shareable and more reusable as well as standard reusable code to aid access and flexibility of knowledge. “We use small pieces of open source code especially when it is simple and without unwanted features. Ultimately it must be something we can (if needed) maintain ourselves without typically any business support model. We are however avid users of open standards especially those regarding knowledge handling and manipulation.” Chapman believes that much open source code, which could be useful in the enterprise, can cause problems due to the nature of the licenses being incompatible with an organizations business model (i.e. as is frequently the case, a closed source solution). “For this reason much open source code in such cases must be avoided to the point where K-Now are forced to recode proprietary solutions.” Chapman says. “On the interfaces and semantic knowledge handling, we use open source,” he added. He referenced two players of interest who are leveraging open source in the communities he works: Talis, and Aduna. “But is the consultancy and support business model scalable? Not really,” he asserts.
K-Now having spun out of Sheffield University is closely connected to the Sheffield University led X-Media Project. This project is one of the largest European Union funded IT research projects (multi-million euro and industry involvement) combining 15 partners from UK, Germany, Italy, France, Slovenia, Greece and Norway. This project continues to develop standards and technology for knowledge management using semantic technology in large organizations expecting to impact industry operations well into 2015 (with launch in 2006), Chapman contends.
Chapman points out that perhaps the biggest industry currently interested in semantic technologies is the pharmaceutical industry. One suggested reason is that the companies are younger, they have less legacy data issues to contend with, and quicker adoption of new technologies, he says. Relationships between labs across the world and improved collaboration are hugely impactful for science. At the end of the day, overcoming the issue of sharing data with semantic technologies remains a security question. Semantic IT infrastructure frameworks are about making it easier to share knowledge. Not every corporation is up for that outside of their own departments or across departments or even in industry discussions. sharing knowledge only where desired. Chapman asserts that without this there will be little adoption.
Comments
One Response to “K-Now: Here and Now”
-
- Subscribe to Beyond Search
Feature archive
News archive
-
Interesting article; I’d have to disagree with some of Chapman’s points on open source though. Incompatible licensing can be a problem if you want to take open source code and build it into a closed source project, but of course that’s kind of the point of open source, that it doesn’t get hidden away in a closed wrapper. There are licenses that allow interopability of course – for example Lucene’s.
As to business models, it’s perfectly possible to build a business on consultancy and support – we’re doing it here – and there are much larger examples (Redhat springs to mind). It sounds like Chapman just hasn’t found the right people to work with!