Future Work Notes

• Generic framework from IJCAI'19 submission ==> journal version?
• Local explanation in iterative planning, question "why not phi?" relative to current candidate plan pi: return only those properties psi entailed by phi where pi not models psi, ie those psi that would necessarily become true were phi enforced. In OSP goal exclusion context: those neg AND B where pi makes AND B true. Further simplifies computation (set as soft goals only those b that pi makes true), and smaller explanation size. Note done in AAAI'20 submission ==> journal version?