Legal Case Winning Party Prediction With Domain Specific Auxiliary Models

Legal Case Winning Party Prediction With Domain Specific Auxiliary Models Sahan Jayasinghe, Lakith Rambukkanage, Ashan Silva, Nisansa de Silva, Amal Shehan Perera Department of Computer Science & Engineering, University of Moratuwa, Sri Lanka {sahanjayasinghe.17, rambukkanage.17, ashansilva.17, nisansaDds, shehan}@cse.mrt.ac.lk Abstract Sifting through hundreds of old case doc- uments to obtain information pertinent to the case in hand has been a major part of the legal profession for centuries. However, with the expansion of court sys- tems and the compounding nature of case law, this task has become more and more intractable with time and resource con- straints. Thus automation by Natural Lan- guage Processing presents itself as a viable solution. In this paper, we discuss a novel approach for predicting the winning party of a current court case by training an an- alytical model on a corpus of prior court cases which is then run on the prepared text on the current court case. This will allow legal professionals to efficiently and precisely prepare their cases to maximize the chance of victory. The model is built with and experimented using legal domain specific sub-models to provide more visi- bility to the final model, along with other variations. We show that our model with critical sentence annotation with a trans- former encoder using RoBERTa based sen- tence embedding is able to obtain an accu- racy of 75.75%, outperforming other mod- els. Keywords: Natural Language Process- ing, Legal Domain, Case Law, Transformer En- coders 1 Introduction Natural Language Processing (NLP) is under- going rapid development and has proven to be practically useful across many text rich do- mains. With the proper utilization of tools and technologies, effective methodologies can be derived to tackle various problems that are repetitive, cognitively demanding and time consuming otherwise. Legal domain is such a text rich domain with a growing need for task automation. Legal domain corpora con- sists of statutes, regulations, constitutions and case law documents among many others which have to be repeatedly and constantly sifted through by legal professionals to obtain infor- mation pertinent to their current case. This research is primarily carried on Case Law doc- uments where a model is train on a corpus of existing case law documents so that a predic- tion of the winning party in a current case law document can be obtained. 1.1 Case Law In the legal domain, when confronted with a new case, where statues, regulations and con- stitutions cannot be used to straightforwardly arrive at a case decision, the courts refer to Case Law. Case Law is the practice of using the information and verdicts of previous cases as arguments for the case in hand where the older cases bear some semblance in one aspect or another to the contemporary case. (Cornell Law School, 2020a). Since case law documents have a predictive, or rather a prescriptive value, in the domain itself, they are valuable resources for predic- tive tasks in both research and practical ap- plications. As time goes on and more and more cases …

Content truncated to 500 words