Natural Language Processing
The NLP Research Unit develops computational models of human languages, focusing on written texts. We are active in the following areas: text mining (document classification, information extraction and ontology population from text, semantic inferences, analysis of the sentiment and of the emotional content of texts); conversational agents (task oriented dialogue systems, question answering, generation of persuasive messages); and development of linguistic resources, particularly for the Italian language.
Contribution to Research
Our research is anchored in Artificial Intelligence and Computational Linguistics. We draw on disciplines such as linguistics, computer science, machine learning, and cognitive science, and aim at developing computational models of language phenomena taking advantage of both statistical (e.g. neural networks) and knowledge-based (e.g. ontologies) approaches to develop state-of-the-art technology.
Contribution to Impact
We aim to converge with the increasing demand of NLP applications in the global market. As an example, in the Health and Well-Being sector, we are developing both text mining systems to extract metadata from clinical reports and a chat-bot for a virtual coach recommending healthy life styles to patients.
We are very proud to contribute to developing tools and linguistic resources for the Italian language, and we distribute most of them freely for research purposes.
Through various scientific collaborations, we launched initiatives such as Recognising Textual Entailment, we developed widely used resources for lexical semantics such as MultiWordNet, WordNet Domains and WordNet Affect, we developed TextPro, a suite of text mining components for various languages, and we contributed to raising interest around the computational study of creative language. We are among the founders of the Italian Association for Computational Linguistics (AILC) and we are regularly involved in the scientific community, chairing shared tasks (e.g. EVALITA) and conferences (e.g. CLIC-it 2014 and AI*IA 2018). We have organised the ACL 2019 conference in Florence and the workshop series on NLP and Journalism and Health Text Mining and Information Analysis.