Dr. Ben Medlock

Co-founder, SwiftKey

As Co-founder and CTO of SwiftKey, Ben invented an intelligent typing system for smartphones that has transformed the text input industry. The software has become the global best selling app on Android and is preinstalled on tens of millions of mobile devices across the world. Based on advanced natural language processing (NLP) and machine learning, the system adapts to individual language usage and device interaction, making touchscreen typing fast and enjoyable. SwiftKey won "most innovative mobile app" at the 2012 GSMA awards, and the company was ranked as the world's 6th most innovative mobile business by Fast Company. Ben developed his expertise in NLP and machine learning during eight years of computer science research, which culminated in a PhD from the University of Cambridge. He has reviewed for a number of prominent international journals, and his academic work is published in ACL, the leading conference for NLP research.
Ben - photo

Research Interests

  • Computer Science
  • Natural Language Processing
  • Applied Machine Learning

  • Projects and Papers

    • Investigating Classification for Natural Language Processing Tasks, VDM 2008  (Amazon)
    • A New Dataset and Method for Automatically Grading ESOL Texts, ACL 2011 (pdf)
    • Exploring Hedge Identification in Biomedical Literature, JBMI 2008 (Elsevier)
    • Weakly Supervised Learning for Hedge Classification in Scientific Literature, ACL 2007 (pdf)
    • An Adaptive Language Model Approach to Spam Filtering on a New Corpus (GenSpam), CEAS 2006 (pdf)
    • An Introduction to NLP-based Textual Anonymisation, LREC 2006 (pdf)
    • Paper on LM-based Spam Filtering (LingSpam) - 2003 (pdf)
    • Masters Thesis - A Generative, Adaptive LM Approach to Spam Filtering (pdf)
    • Bachelors Thesis - A Tool for GLR Parsing in Haskell (pdf)

    Other Links


    • Download the GenSpam email corpus here (download page).
    • Download the ITAC anonymisation corpus here (download page).
    • Download the Hedge Classification dataset here (download page).