This article describes the Vaccine/Injury Project Corpus, a collection of legal decisions awarding or denying compensation for health injuries allegedly due to vaccinations, together with models of the logical structure of the reasoning of the factfinders in those cases. This unique corpus provides useful data for formal and informal logic theory, for natural-language research in linguistics, and for artificial intelligence research. More importantly, the article discusses lessons learned from developing protocols for manually extracting the logical structure and generating the logic (...) models. It identifies sub-tasks in the extraction process, discusses challenges to automation, and provides insights into possible solutions for automation. In particular, the framework and strategies developed here, together with the corpus data, should allow “top–down” and contextual approaches to automation, which can supplement “bottom-up” linguistic approaches. Illustrations throughout the article use examples drawn from the Corpus. (shrink)

The first issue of Artificial Intelligence and Law journal was published in 1992. This paper provides commentaries on nine significant papers drawn from the Journal’s second decade. Four of the papers relate to reasoning with legal cases, introducing contextual considerations, predicting outcomes on the basis of natural language descriptions of the cases, comparing different ways of representing cases, and formalising precedential reasoning. One introduces a method of analysing arguments that was to become very widely used in AI and Law, namely (...) argumentation schemes. Two relate to ontologies for the representation of legal concepts and two take advantage of the increasing availability of legal corpora in this decade, to automate document summarisation and for the mining of arguments. (shrink)

Sentence boundary detection (SBD) represents an important first step in natural language processing since accurately identifying sentence boundaries significantly impacts downstream applications. Nevertheless, detecting sentence boundaries within legal texts poses a unique and challenging problem due to their distinct structural and linguistic features. Our approach utilizes deep learning models to leverage delimiter and surrounding context information as input, enabling precise detection of sentence boundaries in English legal texts. We evaluate various deep learning models, including domain-specific transformer models like LegalBERT and (...) CaseLawBERT. To assess the efficacy of our deep learning models, we compare them with a state-of-the-art domain-specific statistical conditional random field (CRF) model. After considering model size, F1-score, and inference time, we identify the Convolutional Neural Network Model (CNN) as the top-performing deep learning model. To further enhance performance, we integrate the features of the CNN model into the subsequent CRF model, creating a hybrid architecture that combines the strengths of both models. Our experiments demonstrate that the hybrid model outperforms the baseline model, achieving a 4% improvement in the F1-score. Additional experiments showcase the superiority of the hybrid model over SBD open-source libraries when confronted with an out-of-domain test set. These findings underscore the importance of efficient SBD in legal texts and emphasize the advantages of employing deep learning models and hybrid architectures to achieve optimal performance. (shrink)

In this work we study, design, and evaluate computational methods to support interpretation of statutory terms. We propose a novel task of discovering sentences for argumentation about the meaning of statutory terms. The task models the analysis of past treatment of statutory terms, an exercise lawyers routinely perform using a combination of manual and computational approaches. We treat the discovery of sentences as a special case of ad hoc document retrieval. The specifics include retrieval of short texts, specialized document types, (...) and, above all, the unique definition of document relevance provided in detailed annotation guidelines. To support our experiments we assembled a data set comprising 42 queries which we plan to release to the public in the near future in order to support further research. Most importantly, we investigate the feasibility of developing a system that responds to a query with a list of sentences that mention the term in a way that is useful for understanding and elaborating its meaning. This is accomplished by a systematic assessment of different features that model the sentences’ usefulness for interpretation. We combine features into a compound measure that accounts for multiple aspects. The definition of the task, the assembly of the data set, and the detailed task analysis provide a solid foundation for employing a learning-to-rank approach. (shrink)