Description
The BSC's Health Language Technology group generates AI and deep-learning based natural language processing resources, including components, annotated data/protocols. It also defines evaluation scenarios to assess and monitor the quality of implementations by the global research community through high impact open data benchmark shared tasks (BioCreative/IberLEF/CLEF). Our text mining technologies serve to exploit unstructured data types that are of practical relevance for data analytics and predictive modelling applications that use literature data, clinical records, social media and patents, covering content in English, Spanish and Catalan.
Our main research application scenarios cover HPC-empowered text mining systems developed for use-cases related to occupational health, cancer (incl. comorbidities and tumour morphology), COVID-19, rare diseases as well as gene regulatory networks, biomaterials and chemical entity information extraction, drug-protein target literature mining.
