Zero-shot Transfer of Article-aware Legal Outcome Classification for European Court of Human Rights Cases

Holzenberger et al. 2020 has modeled statutory Legal Judgment Prediction (LJP) has recently reasoning by classifying US tax law provisions gained considerable attention in the mainstream concatenated with textual case descriptions. We NLP community (e.g., Aletras et al. 2016; build on this prior work in two ways. First, we Chalkidis et al. 2019, 2021, 2022b; Santosh et al. develop and evaluate our model on a public dataset 2022, 2023). In LJP, the outcome of a case should (Chalkidis et al., 2022b) of cases by the European be classified/predicted based on a textual description Court of Human Rights (ECtHR), which hears complaints of case facts. In actual legal reasoning, legal by individuals about possible infringements practitioners (e.g., advocates, judges) determine relevant of their rights enshrined in the European Convention rules from the sources of law (e.g., statutes, on Human Rights (ECHR) by states. To the regulations, precedent) that are relevant to the case best of our knowledge, this is the first work applying at hand. They then carry out an analysis to determine article-aware case outcome prediction setting which rules apply to the case at hand, and to human rights adjudication.