Invited talks programme

All invited talks are in the Middleton Room, from 16:15-17:15 (Monday 16:30 - 17:30). They will consist of a 45 min talk with 15 min for discussion. The invited talks are free and open to all participants.

This page will be updated with more details nearer the time.

• Monday 18 June: GATE team.
  
  Title: GATE demos and future previews
  
  
• Tuesday 19 June: Phil Gooch, City University, London.
  
  Title: 'It could be lupus': Identifying narrative event chains in clinical notes
  
  Abstract: Clinical reports such as progress notes, discharge summaries and pathology reports contain information essential for treatment planning, decision support, research and clinical audit. However, extracting explicit named entities such as people, symptoms, conditions, tests and treatments is insufficient for this purpose. Resolution of coreference - where two expressions refer to the same real-world entity or event - is required to uncover linked narrative events that are often implicit in the text. This problem has only recently been considered for clinical notes, despite being a well-researched topic for general newswire and media texts. In this presentation, we consider what Huggy Bear and House can tell us about coreference relations that are easily expressed and understood by humans, but are difficult to process computationally.
  
  Slides: It could be lupus
  
  

• Wednesday 20 June: Antony Scerri, Elsevier.
  
  Title: Text Analytics at Elsevier
  
  Abstract: With our large archive of published literature we are always looking at different ways to use it. GATE has enabled us to apply NLP techniques quickly and easily to help with developing ideas. We will summarise a few projects where we have used GATE to help get things rolling. We will also present our SciVerse platform and how we have enabled a simple entity recognition service. Finally we'll talk about our latest efforts looking at applying statistical NLP techniques across our entire collection.
  
  Slides:
  
  

• Thursday 21 June:Tom Storrar, The National Archives, London.
  
  Title: Making the UK Government Web Archive More Usable
  
  Abstract: The UK Government Web Archive is one of the largest and most used web archives in the world. It is a high quality web archive, with over a billion pages and documents, across different times and many different government websites. This means that conventional search approaches are not as effective as our users need. Driven by this and by the desire to unlock specialist domain knowledge, the Semantic Knowledge Base was built as a solution. GATE is used as our NLP tool, while ontologies and linked data provide users with contextual information and allow inference. This presentation will be about how we developed GATE and Mimir applications with Sheffield to meet our very specific needs.
  
  Slides: