domain-independent knowledge

Domain-independent knowledge is the kind of common sense or general knowledge that one has without studying a specific topic or area of expertise. Simple domain-independent knowledge, such as the idea of generalising from past examples, may be implicitly built into the way algoritms work. More extensive domain-independent knowledge, such as the fact that cats chase birds, is far harder for machines, yet simple for humans. The long standing CYC Project project is one attempt to deal with this and one can argue that large language models do capture some level of general knowledge albiet implicitly in the deep neural network weights. The term domain-independent knowledge is used in contrast with domain-dependent knowledge, the specialised knowledge of a specific area.