Epistemic Interaction

Alan Dix

Computational Foundry, Swansea University, Wales

Draft paper for comment 1 November 2020

Download full paper (PDF, 186K)

There is a growing expectation that forms of machine learning or other data-driven algorithms will enter virtually every aspect of human–computer interactions.  In many examples the user’s interactions in the physical or digital world are monitored in order to modify subsequent system behaviour.  Usually this is done with minimal or no change to the primary interactions, you act as normal, but these actions are passively sensed and analysed.  However, there is the potential for minor shifts in interaction that can substantially increase the information available for system adaptation.  That is, we can deliberately design interactions to increase the knowledge available to the system – I call this epistemic interaction.











Alan Dix 1/11/2020