Review for NeurIPS paper: Active Structure Learning of Causal DAGs via Directed Clique Trees

They lower bound the minimal number of interventions required to orient any graph in terms of the size of the largest cliques in the essential graph of the causal model.