2017/11/15 by Patrick Rodler, Rodler, Patrick, Wolfgang Schmid +3 · 1 citation
Computer Science · Engineering · #Artificial Intelligence (cs.AI) #Bayesian Modeling and Causal Inference #FOS: Computer and information sciences #Fault Detection and Control Systems #Software Testing and Debugging Techniques
paper · pdf · doi:10.48550/arxiv.1711.05508
openalex publication_date 2017/11/15 · openalex created_date 2022/08/31 · openalex updated_date 2026/07/28
Model-Based Diagnosis deals with the identification of the real cause of a\nsystem's malfunction based on a formal system model and observations of the\nsystem behavior. When a malfunction is detected, there is usually not enough\ninformation available to pinpoint the real cause and one needs to discriminate\nbetween multiple fault hypotheses (called diagnoses). To this end, Sequential\nDiagnosis approaches ask an oracle for additional system measurements.\n This work presents strategies for (optimal) measurement selection in\nmodel-based sequential diagnosis. In particular, assuming a set of leading\ndiagnoses being given, we show how queries (sets of measurements) can be\ncomputed and optimized along two dimensions: expected number of queries and\ncost per query. By means of a suitable decoupling of two optimizations and a\nclever search space reduction the computations are done without any inference\nengine calls. For the full search space, we give a method requiring only a\npolynomial number of inferences and show how query properties can be guaranteed\nwhich existing methods do not provide. Evaluation results using real-world\nproblems indicate that the new method computes (virtually) optimal queries\ninstantly independently of the size and complexity of the considered diagnosis\nproblems and outperforms equally general methods not exploiting the proposed\ntheory by orders of magnitude.\n