Natural Language Processing

The NLP team is responsible for making sense of the growing literature using neural NLP methods which increase the speed and reliability of knowledge discovery. 

The team develops neural information extraction methods to automatically extract and manipulate information from distributed, heterogeneous and unstructured textual collections to yield structured and actionable knowledge. Our areas of application include semantic search, association mining from multiple information resources, summarization, automated pathway curation, knowledge graph construction, biological modeling, and their evaluation by domain experts.  We create the foundations of bridging text with knowledge, linking deep learning with deep semantics, enabling the interpretation of scientific text using heterogeneous sources of information.

Our team is closely linked with the pioneering research of the National Centre for Text Mining developing NLP tools, resources and interoperable infrastructure. Our current research topics include named entity recognition and normalization, coreference resolution, relation extraction at document level, unsupervised relation extraction, event extraction, development of labeled data using crowdsourcing and active and proactive methods, scientific citation extraction, modality and temporality detection.