AI driven textual data generation, analysis and information extraction is invaluable in almost every domain. In HR it can be used to automatically extract keywords, in healthcare, to scan scientific literature for compound-adverse reaction pairs, and chatbots or conversational agents are perhaps the hot topic of 2023. Beyond this, the notion of text can be generalized as a sequence of symbols that have contextual meaning, opening the door for even more valuable applications, such as protein language models for generating meaningful biological sequences. IDLab focuses on:
Specific research track differentiators include a focus on causality in collaboration with Stanford University, to support explainable AI, domain-tailored solutions, such as NLP for healthcare applications, both traditional human text and protein language applications, using the latest neural network architectures (e.g., transformers in BERT-like models) and the creation of new, valuable training data sets.
IDLAB Ghent has expertise in (applied) research on machine learning for textual data, mainly within the domains of:
Healthcare
e.g. extraction of adverse reactions from medical literature and real-world data, protein language modelling
HR
e.g. automatic processing of vacancies/CVs, … (e.g., skill extraction and matching)
Education
e.g. automatic question generation (factual questions; language learning exercises; distractor generation for multiple choice)
Media
e.g. Classification of news articles
