RELATED

1st International Workshop

Relations in the Legal Domain

@ICAIL 2021 – The 18th International Conference on Artificial Intelligence and Law, University of São Paulo, Monday June 21 – Friday June 25.

The RELATED workshop focuses on the wide-ranging topic of relations in the legal field. Typical goals in relations studies are: ranking of normative provisions and judgments; classification of implicit and explicit links between norms; handling potential or actual conflict between technical and goal-oriented norms with recourse to mechanisms such as defeasibility and balance of principles; discovery of synonyms and near-synonyms in documents from different languages and contexts (jurisdictions, document type); discovery of instances of count-as terms and policies to facilitate analogical reasoning and compliance checking. Relations are an important aspect of many sub-disciplines of AI & Law – ontologies, logic, argumentation and network analysis. This workshop is intended to be a forum bringing together all these sub-disciplines to focus on relations from different perspectives.

The challenging task of automatically identifying relations between concepts expressed in textual documents has traditionally been addressed by NLP techniques. Once such relations are identified in the text, they can be formalized to become part of ontological models in a certain legal domain of interest. The extraction of relations between legal concepts and entities is even more pertinent when considering the increasing relevance of knowledge graphs in the Semantic Web context. For instance, applications that extract relations between norms, legal concepts and documents may look at the Linked Open Data (LOD) cloud for reusing, integrating and expanding existing data on the Web.

Navigating the interplay between legal text fragments beyond terms, such as normative provisions, principles, policies, arguments, is a challenging aspect of legal reasoning. There is a rich theoretical literature on relations between norms, principles and factors. Moreover, recent years have seen important developments in the field of network analysis, particularly the identification, classification and weighting of legal citations in order to build a network of related legal sources and rank their individual importance. Less explored is the development of systems for automated identification of implicit relations between norms in terms of, for instance, conceptual similarity, motivation and conflict. Also of relevance to this workshop are topic networks, which can be useful not only for semantic grouping of legal text fragments, but also in comparing different legal documents, such as legislative texts, judgments, national strategy documents, guidelines issued by institutional bodies, and policy documents issued by companies. Contributions are welcome from NLP, machine learning, logic and argumentation.