Graph weaknesses in commonsense causal representation
Lawrence J. Mazlack · Annual Conference on Computers · 2009
Causal reasoning occupies a central position in human reasoning. In order to algorithmically consider causal relations, the relations must be placed into a representation that supports manipulation. The most widespread causal representation in current usage is directed acyclic graphs. However, they are severely limited in what portion of the every day world they can represent. Some of the required Markov conditions do not fit with commonsense reasoning. More importantly, cycles must be represented and they cannot be represented in acyclic graphs. Additionally, shifts in grain size are overly limited. Commonsense understanding deals with imprecision, uncertainty and imperfect knowledge. An algorithmic way of handling and representing causal imprecision that includes cycles is needed.