The opportunities and threats posed by machine learning were debated during a lecture by serial entrepreneur Hermann Hauser at Alliance MBS.
Dr Hauser, co-founder of a string of high-tech companies including Acorn Computers and of venture capital firm Amadeus Capital Partners, said machine learning was a rare horizontal technology that could be applied right across society. “It can be applied to most, if not all, sectors. It changes everything and creates new business models.”
He cited the example of the automotive industry where machine learning is the key to autonomous driving. “Here is an industry which is being impacted by two spectacular disruptions. One is the electric motor, and the other is autonomous driving.”
Dr Hauser said we had in recent times seen huge advances in the use of machine learning – defined as systems that learn data rather than follow pre-programmed rules – such as through clever speech and face recognition.
Looking ahead he stressed that at the heart of machine learning was probability. “If you want to describe the real world it is much better to associate probability to all things you observe. It gives you a more powerful way of predicting what happens next.”
However he conceded there were major societal and ethical challenges ahead. For instance he said there was a potential threat to more than half of jobs from machine learning, while another major challenge was that machine learning needed human goals.
Steve Furber, ICL Professor of Computer Engineering at the University of Manchester, said deep learning will bring many changes but we were only just at the beginning of the journey. “Deep learning has only just begun, these are the first steps. What we have heard about tonight are lots of very exciting niche applications which have solved some interesting problems. But there has been no progress towards understanding the nature of human intelligence or towards being able to emulate it in our machines.”
Welcoming Dr Hauser to Alliance MBS, Professor Jakob Edler, Executive Director of the Manchester Institute of Innovation Research, said this was a topic on everyone’s minds. “Whatever we do every day seems to now be helped by algorithms. Sometimes it seems that search engines know better what we want than we do ourselves. It is fascinating, but worrying too. I’m not sure if I want machines to learn so much about me. There is a balance to be struck.”