Eur. Phys. J. Special Topics 214, 461-479 (2012)
https://doi.org/10.1140/epjst/e2012-01702-4

Regular Article

Causality discovery technology

¹ Oxford e-Research Centre, University of Oxford, UK
² Visualization and Interactive Systems Institute, University of Stuttgart, Germany
³ Department of Computer Science, University of Oxford, UK
⁴ Department of Engineering Science, University of Oxford, UK

^a e-mail: min.chen@oerc.ox.ac.uk

Received: 1 August 2012
Revised: 9 October 2012
Published online: 5 December 2012

Abstract

Causality is the fabric of our dynamic world. We all make frequent attempts to reason causation relationships of everyday events (e.g., what was the cause of my headache, or what has upset Alice?). We attempt to manage causality all the time through planning and scheduling. The greatest scientific discoveries are usually about causality (e.g., Newton found the cause for an apple to fall, and Darwin discovered natural selection). Meanwhile, we continue to seek a comprehensive understanding about the causes of numerous complex phenomena, such as social divisions, economic crisis, global warming, home-grown terrorism, etc. Humans analyse and reason causality based on observation, experimentation and acquired a priori knowledge. Today’s technologies enable us to make observations and carry out experiments in an unprecedented scale that has created data mountains everywhere. Whereas there are exciting opportunities to discover new causation relationships, there are also unparalleled challenges to benefit from such data mountains. In this article, we present a case for developing a new piece of ICT, called Causality Discovery Technology. We reason about the necessity, feasibility and potential impact of such a technology.