Decomposable Negation Normal Form

Decomposable Negation Normal Form (d-DNNF) is a knowledge compilation technique that transforms Boolean formulas into a form enabling efficient probabilistic inference and model counting. Current research focuses on optimizing d-DNNF representations, including developing algorithms to prune redundant structures and exploring the limitations of d-DNNF, particularly concerning negation and its impact on succinctness compared to other models like SDDs. These advancements improve the efficiency of various applications, such as probabilistic reasoning and explainable AI, by enabling faster computation of tasks like model enumeration and top-k model selection.