Skip to Content

knowledge advance col-laboratory

The main idea of knowledge advance col-laboratory is that a user can extend their knowledge by browsing his autonomously expanding knowledge network. This is a way for people becomes to be a master of knowledge. The autonomously expanding are implemented with several AI agents. They take in charge of several aspects of users' learning paradigm. First aspect is learning, or regarding as context awareness. The agent has the capacity to know where the user is and what he is dedicating in. Meanwhile, the accumulating user activities would form a learning path which help agents to discover his/her knowledge favorite. Second aspect is knowledge filtering. Knowledge is filtered and also integrated by pre-defined goal or desire-derived goal. Then the small knowledge knowledge network can be formed into a bigger one. This action can also be applied on user. User can automatically form a group by their interest. Third aspect is in collaboration. This would form a new view. It has two sides. One is work-flow oriented and the other One is aimless walking. The prior is easy to understand. The collaboration is started in a format which might be predefined or revisable. The second would be a challenge. It highly depends on the result from second aspect.

Vijjana works with AI agents, and facilitates user' knowledge acquisition. The basic knowledge unit JAN in Vijjana is based on our approach to IEEE standard 1484.12.1. This has become the core part of the Vijjana model. JANs are contributed by users and agents work upon them based on users' interest that may change in time.

The model is depicted as
Vijjana-X = { J, T, R | dA, oA, cA, vA, sA, rA}, where
X = the domain name,
J = the collection of Jans in the Vijjana-X,
T = the Taxonomy OR pattern set used for classification of Jans,
R = the domain specific relations;
dA = the discovery agent which find relevant Jans,
oA = the organizing agent which interlinks Jans based on R,
cA = the consistency/completeness agent,
vA = the visualization agent,
sA = the search agent,
rA = the rating agent.

In the Vijjana model, discovery agent (dA) helps user to gain new knowledge units through filtering a large amount of source information, at the same time it performs key feature extraction on learning user\u2019s preference. Organizing agent (oA) and Consistency Agent (cA) are responsible for internal system management. Amusingly, Darwin\u2019s theory of evolution is also seen to be working in knowledge management lifecycle, where new updated knowledge would replace the older one, high frequently hit information should be cached and low profiled information should be categorically assembled. Also, oA and cA ensure that the model becomes self-organizable. Visualization Agent (vA) construct the knowledge network from a user's personal perspective or any global views and display them in several structured format such as radial view, shown as Fig 1., or tree view. The Search Agent (sA) would filter information according to specified criteria defined through user requirement or reasoning approaches. Rating Agent (rA) combine with knowledge discovery agent dA , consistency/completeness agent cA and organizing agent oA,, and form a knowledge acquisition cycle in the model. Comments from users are also useful sources of features extraction during pattern recognition and classification. A pattern-relearning procedure is invoked within rA to consolidate previous results or recall oA and cA to perform desired newer actions.