Learning Qualitative Causal Models via Generalization & Quantity Analysis

Learning causal models is a central problem of qualitative reasoning. We describe a simulation of learning causal models from exemplars that uses progressive alignment and qualitative process theory to derive plausible qualitative causal models from observations. We show how protohistories can be created via progressive alignment and used to infer causality. The result, a causal corpus, can make simple predictions and set the stage for more sophisticated qualitative models. The simulation has been successfully tested with learning causal mechanisms of three physical scenarios, with encouraging results.