Probabilistic reasoning in intelligent systems: networks of plausible inference
Probabilistic reasoning in intelligent systems: networks of plausible inference
The causal Markov condition, fact or artifact?
ACM SIGART Bulletin
A qualitative Markov assumption and its implications for belief change
UAI'96 Proceedings of the Twelfth international conference on Uncertainty in artificial intelligence
Probabilistic temporal networks: A unified framework for reasoning with time and uncertainty
International Journal of Approximate Reasoning
| Hi-index | 0.00 |
The question of what constitutes a causal model is significant, as it will lead to a better understanding of the role of causality and its heuristic role in probabilistic reasoning. It has been argued that in certain cases domain specific considerations can be appealed to in the construction of more efficient causal models that are "nonstandard" in the way networks reflect anomalous correlations between nodes. For the most part, the causal assumptions generally invoked [Pearl 1988], [Spirtes, Glymour, Scheines 1993] do lead to systematic efficiencies based on a reduction in computational requirements for the models they produce. The goal of this paper is to uncover some of the assumptions about causality that undergird current causal models, assumptions which should be kept in mind by those invoking causal relations in the construction of discovery algorithms for causal networks.