Machine learning: the power and promise of computers that learn by example

The report touches on how this area is developing and the societal repercussions of using machine learning in healthcare, retail, law and banking. The UK has a key competitive advantage in this area, as many of the early innovations in the field were due to UK researchers. Looking towards the future, with the aggressive acquisition of university spin-offs and hiring of machine learning academics in industry a number of steps are required to keep the UK competitive academically.

Sofia Olhede notes: "Being involved in this report has been very interesting. As part of the work, we explored UK public opinions in this field. It was very encouraging to meet so much enthusiasm in terms of the potential gains of using machine learning technologies in practical applications, but this was tempered by very rational fears of what could happen if humans are not kept in the decision-making loop. Key challenges for the future are making sure machine learning technologies develop that are both fair and transparent."

The Royal Society committee's report is part of a larger international drive to fully understand the impacts of using algorithms in a semi-automated fashion in society. In Germany, the Leopoldina National Academy of Science has a working group on how volumes of data are treated, as in the US, the Institute of Electrical and Electronics Engineers have launched an initiative in Ethical Considerations in the Design of Autonomous Systems.
 

Links

Leopoldina National Academy of Science
Royal Society - Machine Learning
Institute of Electrical and Electronics Engineers
Article on Algorithm Fairness and Transparency by Olhede & Rodrigues

Source

Article on Royal Society website