do not care that someone else imagine this one differently ... if i have data base - i have data base and i will not put there = mark just to say yes i got it like file system....in fact the difference exists because before - some time ago knowledge base was rules+facts=domain knowledge, and it was stored in some data files...as well domain knowledge mean to be knowledge about one specific case, system - now it is quite changed, because there exists not relationship: 1 expert - 1 knowledge base, but: many experts - 1 knowledge base and knowledge base can store knowledge about more than one system or case...so it means have changed knowledge base parts ... knowledge base is constructed as data structure collection+rules (besides data structure contains knowledge how system stores and manage knowledge) which purpose is to acquire, store and compute knowledge - so knowledge base is now as kind of container that can be abstract (from the point of view it is not one strictly separated component in systems architecture), but domain knowledge that is later provided by experts is represented and stored using knowledge base constructs (used data structures links represented - to store it in the way it can be interpreted and understood as knowledge)
so how, why ... first there were data files where the data was stored (still remember my first experiment - data file where was stored some expert`s knowledge about system - everything worked based on frame hierarchy...), then later data bases (still data files, but with a specific structured organisation - which changed view) were created but still as knowledge base mean to be consisting of facts +rules (mostly mean if then format) although both are kind of bases the meaning, purpose is different ...
in some papers data bases are simply called knowledge base...i do not like that because in fact even if in system architecture the data base is used to store data, information, facts...this do not store knowledge - for example - without discovering you cant get complete understanding about represented system....so therefore now exists trend ...Knowledge Discovery in Data bases (KDD) so how and where the knowledge appears...it is discovered and represented (after it is stored) using knowledge representation schema...and to complete mentioned actions there are rules, logic, functions used
what i did not mentioned before is inference machine...that was often mentioned as component of knowledge based/ expert systems ....because Intelligent systems are created with purpose to reason, adapt, etc... this component is significant as well - but now it is usually created using different application functions,that are connected with queries (have seen that search engine and inference engine are mentioned as separated parts, what is important them can be strongly connected and improve ea.)
the same as before were teaching just programming language and now teach builders...and it is difference...
the same now rules and logic, the same as data structures and domain knowledge have templates /patterns...
any structure puts some constraints :-)
No comments:
Post a Comment