Define a set of Named Entity tags (Person, Organisation, Location, GPE, diseases, etc.), detect and highlight the entities mentioned in the text.
Classify the entire text as negative, positive, or neutral and explain how the classifier decided (visualising attention)
Detect the polarity of the text towards an entity (which may be part of the text or not) and explain how the classifier decided.
Analyse large document collections, then process and sort this information by applying probabilistic models to extract hidden topics. Discover the main topics talked about.
Check some of the solutions (non-exhaustive) we have developed and used in various settings.
Given a text and an innovation taxonomy (optional), segment the text into sections and classify every sentence – if it conveys an innovation statement and what type of innovation
Analyse scientific publications and policy documents in the health domain and identify topics in rare diseases.