Tensor-Based Approach to Diagnostic Causal Network Modeling

In the context of diagnostic and abduction, this chapter is built upon the meta-hypothesis stating that the true cause of the observation (referred to as the concept) belongs to a set of explanations. In addition, finding the concept means a strategy of observation planning (referred to as the policy). A tensor-based approach to causal network modeling is first proposed in the context of a new cognitive strategy expressed as a merge of diagnostic and abduction. Next, an overview on the minimum length solution algorithm is given. This chapter also proposes an original example that applies this algorithm. A generalization of the algorithm is suggested. Illustrative simulation results are offered. This chapter is dedicated to Prof. Imre J. Rudas, to celebrate his 75th anniversary.