Solving Organisational Communication Problems Using Machine Reading
Today, much of an organisation's knowledge is stored in natural language. Natural language is hard for computers, so technology hasn't so far been able give immediate and direct access to this knowledge - even with search it can still take too long, or it can be too difficult, to get the answers you need when you need them. Bloomsbury AI aims to solve this problem by making advances in the speed and accuracy with which computers can read documents and answer questions about them. In this talk, they present research on a key question in such methods - whether or not neural question answering systems trained on standard datasets can generalise sufficiently to new domains.
Jonathan Godwin is the Product Manager of Bloomsbury AI. Previously, he was the CEO of a startup, worked as a data scientist and founded Social Finance Digital Labs. He has an MSc in Machine Learning from UCL where he won multiple awards for academic achievement.