The structures of discourse used by legal and ordinary languages share differences that foster technical issues when applying or fine-tuning general-purpose language models for open-domain question answering on legal resources. For example, longer sentences may be preferred in European laws (i.e., Brussels I bis Regulation EU 1215/2012) to reduce potential ambiguities and improve comprehensibility, distracting a language model trained on ordinary English. In this article, we investigate some mechanisms to isolate and capture the discursive patterns of legalese in order to (...) perform zero-shot question answering, i.e., without training on legal documents. Specifically, we use pre-trained open-domain answer retrieval systems and study what happens when changing the type of information to consider for retrieval. Indeed, by selecting only the important parts of discourse (e.g., elementary units of discourse, EDU for short, or abstract representations of meaning, AMR for short), we should be able to help the answer retriever identify the elements of interest. Hence, with this paper, we publish Q4EU, a new evaluation dataset that includes more than 70 questions and 200 answers on 6 different European norms, and study what happens to a baseline system when only EDUs or AMRs are used during information retrieval. Our results show that the versions using EDUs are overall the best, leading to state-of-the-art F1, precision, NDCG and MRR scores. (shrink)

Legal text retrieval serves as a key component in a wide range of legal text processing tasks such as legal question answering, legal case entailment, and statute law retrieval. The performance of legal text retrieval depends, to a large extent, on the representation of text, both query and legal documents. Based on good representations, a legal text retrieval model can effectively match the query to its relevant documents. Because legal documents often contain long articles and only some parts are relevant (...) to queries, it is quite a challenge for existing models to represent such documents. In this paper, we study the use of attentive neural network-based text representation for statute law document retrieval. We propose a general approach using deep neural networks with attention mechanisms. Based on it, we develop two hierarchical architectures with sparse attention to represent long sentences and articles, and we name them Attentive CNN and Paraformer. The methods are evaluated on datasets of different sizes and characteristics in English, Japanese, and Vietnamese. Experimental results show that: (i) Attentive neural methods substantially outperform non-neural methods in terms of retrieval performance across datasets and languages; (ii) Pretrained transformer-based models achieve better accuracy on small datasets at the cost of high computational complexity while lighter weight Attentive CNN achieves better accuracy on large datasets; and (iii) Our proposed Paraformer outperforms state-of-the-art methods on COLIEE dataset, achieving the highest recall and F2 scores in the top-N retrieval task. (shrink)

Modeling law search and retrieval as prediction problems has recently emerged as a predominant approach in law intelligence. Focusing on the law article retrieval task, we present a deep learning framework named LamBERTa, which is designed for civil-law codes, and specifically trained on the Italian civil code. To our knowledge, this is the first study proposing an advanced approach to law article prediction for the Italian legal system based on a BERT (Bidirectional Encoder Representations from Transformers) learning framework, which has (...) recently attracted increased attention among deep learning approaches, showing outstanding effectiveness in several natural language processing and learning tasks. We define LamBERTa models by fine-tuning an Italian pre-trained BERT on the Italian civil code or its portions, for law article retrieval as a classification task. One key aspect of our LamBERTa framework is that we conceived it to address an extreme classification scenario, which is characterized by a high number of classes, the few-shot learning problem, and the lack of test query benchmarks for Italian legal prediction tasks. To solve such issues, we define different methods for the unsupervised labeling of the law articles, which can in principle be applied to any law article code system. We provide insights into the explainability and interpretability of our LamBERTa models, and we present an extensive experimental analysis over query sets of different type, for single-label as well as multi-label evaluation tasks. Empirical evidence has shown the effectiveness of LamBERTa, and also its superiority against widely used deep-learning text classifiers and a few-shot learner conceived for an attribute-aware prediction task. (shrink)

The first issue of Artificial Intelligence and Law journal was published in 1992. This paper offers some commentaries on papers drawn from the Journal’s third decade. They indicate a major shift within Artificial Intelligence, both generally and in AI and Law: away from symbolic techniques to those based on Machine Learning approaches, especially those based on Natural Language texts rather than feature sets. Eight papers are discussed: two concern the management and use of documents available on the World Wide Web, (...) and six apply machine learning techniques to a variety of legal applications. (shrink)

The contract review process can be a costly and time-consuming task for lawyers and clients alike, requiring significant effort to identify and evaluate the legal implications of individual clauses. To address this challenge, we propose the use of natural language processing techniques, specifically text classification based on deontic tags, to streamline the process. Our research question is whether natural language processing techniques, specifically dense vector embeddings, can help semi-automate the contract review process and reduce time and costs for legal professionals (...) reviewing deontic modalities in contracts. In this study, we create a domain-specific dataset and train both baseline and neural network models for contract sentence classification. This approach offers a more efficient and cost-effective solution for contract review, mimicking the work of a lawyer. Our approach achieves an accuracy of 0.90, showcasing its effectiveness in identifying and evaluating individual contract sentences. (shrink)

Evidences are important proofs to support judicial trials. Automatically extracting evidences from judgement documents can be used to assess the trial quality and support “Intelligent Court”. Current evidence extraction is primarily depended on sequence labelling models. Despite their success, they can only assign a label to a token, which is difficult to recognize nested evidence entities in judgment documents, where a token may belong to several evidences at the same time. In this paper, we present a novel evidence extraction architecture (...) called ATT-MRC, in which extracting evidence entities is formalized as a question answer problem, where all evidence spans are screened out as possible correct answers. Furthermore, to address the data imbalance problem in the judgement documents, we revised the loss function and combined it with a data enhancement technique. Experimental results demonstrate that our model has better performance than related works in evidence extraction. (shrink)

This work aims to explore the application of deep learning-based artificial intelligence technology in sentencing, to promote the reform and innovation of the judicial system. First, the concept and the principles of sentencing are introduced, and the deep learning model of intelligent robot in trials is proposed. According to related concepts, the issues that need to be solved in artificial intelligence sentencing based on deep learning are introduced. The deep learning model is integrated into the intelligent robot system, to assist (...) in the sentencing of cases. Finally, an example is adopted to illustrate the feasibility of the intelligent robot under deep learning in legal sentencing. The results show that the general final trial periods for cases of traffic accidents, copyright information, trademark infringement, copyright protection, and theft are 1,049, 796, 663, 847, and 201 days, respectively; while the final trial period under artificial intelligence evaluation based on the restricted Boltzmann deep learning model is 458, 387, 376, 438, and 247 days, respectively. The accuracy of trials is above 92%, showing a high application value. It can be observed that expect theft cases, the final trial period for others cases has been effectively reduced. The intelligent robot assistance under the restricted Boltzmann deep learning model can shorten the trial period of cases. The deep learning intelligent robot has a certain auxiliary role in legal sentencing, and this outcome provides a theoretical basis for the research of artificial intelligence technology in legal sentencing. (shrink)

If a trademark becomes a generic term, it may be cancelled under trademark law, a process known as genericide. Typically, in genericide cases, consumer surveys are brought into evidence to establish a mark’s semantic status as generic or distinctive. Some drawbacks of surveys are cost, delay, small sample size, lack of reproducibility, and observer bias. Today, however, much discourse involving marks is online. As a potential complement to consumer surveys, therefore, we explore an artificial intelligence approach based chiefly on word (...) embeddings: mathematical models of meaning based on distributional semantics that can be trained on texts selected for jurisdictional and temporal relevance. After identifying two main factors in mark genericness, we first offer a simple screening metric based on the ngram frequency of uncapitalized variants of a mark. We then add two word embedding metrics: one addressing contextual similarity of uncapitalized variants, and one comparing the neighborhood density of marks and known generic terms in a category. For clarity and validation, we illustrate our metrics with examples of genericized, somewhat generic, and distinctive marks such as, respectively, DUMPSTER, DOBRO, and ROLEX. (shrink)

I consider the potential of eight text-scaling methods for the analysis of jurisprudential change. I use a small corpus of well-documented German Federal Constitutional Court opinions on European integration to compare the machine-generated scores to scholarly accounts of the case law and legal expert ratings. Naive Bayes, Word2Vec, Correspondence Analysis and Latent Semantic Analysis appear to perform well. Less convincing are the performance of Wordscores, ML Affinity and lexicon-based sentiment analysis. While both the high-dimensionality of judicial texts and the validation (...) of computer-based jurisprudential estimates pose major methodological challenges, I conclude that automated text-scaling methods hold out great promise for legal research. (shrink)

Court judgments contain valuable information on how statutory laws and past court precedents are interpreted and how the interdependence structure among them evolves in the courtroom. Data-mining the evolving structure of such customs and norms that reflect myriad social values from a large-scale court judgment corpus is an essential task from both the academic and industrial perspectives. In this paper, using data from approximately 110,000 court judgments from Japan spanning the period 1998–2018 from the district to the supreme court level, (...) we propose two tasks that grasp such a structure from court judgments and highlight the strengths and weaknesses of major machine learning models. One is a prediction task based on masked language modeling that connects textual information to legal codes and past court precedents. Another is a dynamic link prediction task where we predict the hidden interdependence structure in the law. We make quantitative and qualitative comparisons among major machine learning models to obtain insights for future developments. (shrink)