Written by Trevor Wood, first published on Cambridge Consultants blog, February 2017
In my favourite episode of Black Mirror, personal data is weaponised. The episode portrays a world where personal information is stored forever and used to determine a ubiquitous social currency. Among the nasty consequences is that a failure to keep up appearances might leave you unable to board an aeroplane or live in a fashionable area.
The stuff of paranoid science fiction for now, perhaps, but Black Mirror gives voice to a well-founded anxiety about the widespread use of personal data. A recent study by Which? highlighted the scale at which personal data could be bought and sold as well as the unscrupulous practices that occur in some parts of the industry. Due care and the appropriate safeguards will be important in the coming years.
There is also reason to look forward with optimism to a future where data about us and our world is available in quantities that were previously unthinkable.
The scale of data available has made possible a class of algorithms which use data to learn how to make classifications and decisions. These algorithms are collectively known as deep learning. For a relatively young technical field, some striking progress has already been made using deep learning techniques to tackle disparate problems.
- Diagnosis of breast cancer with higher accuracy than trained doctors.
- Analysing and captioning images with higher accuracy than humans. Tools using this ability may form a crucial part of systems for self-driving cars.
- Beating a world champion at the famously difficult board game ‘Go’.
One particular application of deep learning evokes the sunnier side of science fiction: the Babel fish from Douglas Adams’s Hitchhiker’s Guide To The Galaxy. (For non-fans, the Babel fish is able to instantly translate between any pair of languages). Deep learning algorithms are approaching the level of translation accuracy of a fluent human for several pairs of languages making automatic real-time translation a possibility in the near-term.
This ought to give us some hope for a future where the glut of data leads to huge benefits for humanity. In the short term, let’s not forget the more modest benefits available today. Netflix’s recommendation system suggests that if I enjoyed Black Mirror then I might like something based on a Douglas Adams novel. It might be onto something.
This article was written by Trevor Wood, Group Leader, Algorithms and Analytics Group at Cambridge Consultants.
Trevor was a guest speaker at Collusion's Artists' Lab, data-culture in January 2017. Cambridge Consultants has joined forces with technology company ARM and arts organisation Collusion in an ambitious two-year national arts programme. Called ‘in_collusion’, it will explore the question of how emerging technologies could change our lives – specifically artificial intelligence, data culture and mixed/augmented reality. Our aim is to combine the skills of our designers and engineers with extraordinary artwork to challenge traditional views of technology and innovation.