YiJia Labs

For years in information technology, experts in the field sacrifice valuable time and energy debugging and fixing database systems. To change this situation, we are developing a brand new knowledge-based system: a system that can be an expert itself.

An expert has comprehensive knowledge in some area of expertise. From reliable advice to the right answer to any question, experts can draw from from a rich bank of memories—instant access through years of work and study—and almost effortlessly synthesize the right move to make from these memories with little more than intuition.

For a system to work as an expert, it needs structured data. Tables with numbers and strings, yes, but also pointers to other objects that in turn have pointers of their own. To represent these systems, we need an object model—an ontology, to borrow the notion from AI literature—with a rich class structure.

To structure this data in this way envisions a cognitive and trait-based approach to data and enables a sort of super memory, a way to instantly store and recall facts from not only bare database entries, but also every intermediate piece of knowledge on the way. This super memory—to recall anything in a blink of an eye—is an major prerequisite to becoming an expert both as a human being and as the database system we envision.

Our knowledge-based system also needs to know facts, relations, and rules about its area of expertise, and it needs to hold and use functions on its own to build new knowledge and information on this area from these sources. This is what makes our systems truly knowledge-based: to be able to create its own data from an intuition of cognitive functions and relationships.

As mentioned above, our current situation began with knowledge as an entirely human affair. As databases develop to be able to store more and more information than ever before, we find ourselves ready to offload routine, knowledge-based tasks to these databases themselves, and we at YiJia Labs are ready to build the systems of automated reasoning and knowledge in these databases to do so.