Argumentation Structure Prediction in CJEU Decisions on Fiscal State Aid
Predicts argument structures in Court of Justice of the European Union decisions concerning fiscal state aid.
Predicts argument structures in Court of Justice of the European Union decisions concerning fiscal state aid.
Combines WordNet and word embeddings for data augmentation in legal text classification.
Introduces a Reddit dataset annotated for sentiment and emotion to support Bitcoin price forecasting.
A study of outcome prediction for Italian value-added tax decisions.
Studies multimodal argument mining in political debates.
Combines offline and online reinforcement learning for energy-management optimization.
Studies argument detection in Court of Justice of the European Union decisions on fiscal state aid.
AMICA is an argumentative search engine for navigating COVID-19 scientific literature.
Introduces SubjectivITA, an Italian newspaper corpus for subjectivity detection.
Presents consumer-facing explanations of potentially unfair contract clauses through the CLAUDETTE tool.