On Semi-Supervised Learning of Legal Semantics

Abstract

In this conference, Professor McCarty reviews his prior work (from 1998, 2007 and 2015) and proposes a research strategy to produce a computational summary of a legal case, which can be scaled up to a realistic legal corpus. The computational challenge was introduced in a 1998 paper on 'Structured Casenotes.' Two steps toward a solution were developed in (i) an ICAIL 2007 paper on 'Deep Semantic Interpretations of Legal Texts,' and (ii) an ICAIL 2015 paper on 'How to Ground a Language for Legal Discourse in a Prototypical Perceptual Semantics.' The current proposal is to combine the 2007 NL model with the 2015 KR model, and to add an ML model based on semi-supervised learning.

Formation d’un dispensateur reconnu aux fins de la formation continue obligatoire du Barreau du Québec pour une durée de 1 heure et 30 minutes. Une attestation de participation représentant 1 heure et 30 minutes de formation sera aussi transmise aux notaires.