Position Paper: LLD is All You Need

tags

Rules as Code

Notes

proper setting for logic programming is intuitionistic logic, not classical logic.

NOTER_PAGE: (2 . 0.29787234042553196)

in intuitionistic logic, we can write universally quantified embedded implications

NOTER_PAGE: (2 . 0.41407528641571195)

in intuitionistic logic we have a proper subset of a first-order language that retains both the definite-answer-substitution property and the goal-directed proof procedure that we want for logic programs.

NOTER_PAGE: (2 . 0.5180032733224222)

it is easy to encode a proof theory for various kinds of legal rules in a logic programming language, such as PROLOG: You get proof search and unification for free in the meta-interpreter.

NOTER_PAGE: (2 . 0.7193126022913258)

Sarah Lawsky argues persuasively in [16] that most statutes are drafted and should be interpreted as rules with exceptions, and she uses Ray Re- iter’s default logic

NOTER_PAGE: (2 . 0.8355155482815058)

a well-drafted statute should never contain unstratified default rules.

NOTER_PAGE: (3 . 0.11783960720130933)

Own can be interpreted as either a predicate or an object, depending on the context, and there is no syntactic distinction between an atomic predicate and an atomic term

NOTER_PAGE: (3 . 0.4402618657937807)

There is now an enormous literature on the “paradoxes” of deontic logic, with no consensus. My view is that the semantics for the deontic modalities is very simple, and the paradoxes arise from complexities in the action language and from our problematic formalisms for default inference.

NOTER_PAGE: (3 . 0.7937806873977087)

justification logic adds the annotation t :P to the proposition P and interprets this com- pound term as “P is justified by reason t.” Essentially, t is a proof of P

NOTER_PAGE: (4 . 0.2037643207855974)

the theory of differential similarity also provides a novel se- mantics for my Language for Legal Discourse.

NOTER_PAGE: (4 . 0.2103109656301146)

broader effort to bridge the gap between machine-learning-based AI and logic-based AI.

NOTER_PAGE: (4 . 0.8502454991816695)