Neuro-symbolic Mixing of Expert Knowledge via Feedback Fuzzy Cognitive Maps

This method uses a Bayes-like theorem to mix directed causal graphs from multiple expert agents or knowledge sources. It defines relevance through probative and material criteria to select expert feedback fuzzy cognitive maps (FCMs) for more accurate neuro-symbolic system predictions.