Text summarization consists in generating a shorter version of an input document, which captures its main ideas. Despite the recent developments in this area, most of the existing techniques have been tested mostly in English and Chinese, due in part to the low availability of datasets in other languages. In addition, experiments have been run mostly on collections of news articles, which could lead to some bias in the research. In this paper, we address both these limitations by creating a dataset for the summarization of legal texts in Portuguese. The dataset, called RulingBR, contains about 10K rulings from the Brazilian Federal Supreme Court. We describe how the dataset was assembled and we also report on the results of standard summarization methods which may serve as a baseline for future works.
展开▼
