Simon Walkden

Former Global Head of End User Engineering at Barclays and leader of Vodafone’s Cloud and Hosting Services, Simon Walkden is steeped in experience at leading international organisations. He has a particular depth and breadth of expertise in ITaaS.

This is embodied in the development, while MD of Flourishing IT Ltd, of an IT Healthcheck service that enables companies to improve delivery, while achieving new levels of internal efficiency and regulatory compliance.

With a passion for reducing cost and complexity, Simon now guides Virtual Clarity’s clients through digital transformation, enabling them to realise the full range of benefits of ITaaS and zero-owned infrastructure.

Connect with Simon.

View articles by Simon

You’ve Got Google MAIL (Machine Learning around Artificial Intelligence)

Recently, Google Cloud sent out their research on Machine Learning. Indeed, we received it by e-mail.

I wonder what a ML application would have made of a few of the excerpts:

* “100% of any company’s future success depends on adopting machine learning”

* “81% of early adopters agree that ML technology can drive down costs”

* “2-5X ROI of most standard ML projects in the first year”

Machine Learning is a structure that allows the machine to observe from the real world, learn and act on learning. Looking around, I’d have difficulty reconciling those statements with companies I see. (OK – I’ve spent the morning trying to sort something out on the UK government website so perhaps things are really less primitive than that recent experience.)

Hopefully, the machine learning tool might say something like “well, the first statement is simply an opinion of a researcher and everyone has opinions about future events which cannot be proved or disproved until later”. I might say that my friend with his organic egg farm is somewhat reliant on the chickens for his future success.

Of the second statement, the application might argue that, by asking early adopters, then you are selecting enthusiasts; and the question was only “can it drive down costs?”, not “has it?”. I’d want to know the real experience of the adopters - which brings us to the last statement. Maybe my learned machine would argue that the numbers have no unit stated and it might be hot air gas increases. My reaction was a dumbfounded “really?”.

For now though, that’s enough cynicism. It is self-evident that if machines can learn and act intelligently in a way similar to human beings, there are seismic implications. That an application can do this is clearly demonstrated in the games field by DeepMind and their AlphaZero application. It learned to play chess in four hours by itself and went on to beat the leading chess program, Stockfish, by means of a bishop sacrifice – a human, irrational action to gamble and win a game. So ML is here, with DeepMind proudly owned by Google. But on the other hand, it is quite easy to baffle Google’s Assistant with a question. At the weekend, we asked the AI-advertised Assistant who was going to win Strictly Come Dancing this year. In another sense, we know ML isn’t quite here, yet.

Google Cloud’s brochure is a useful reminder of the impressive set of tools the company already offers in the AI field – image recognition, voice to text, translation, etc. It also makes the point that cloud is the ideal platform. That combination of tools and on demand resources is powerful. And at the front, the list of fields in which machine learning has begun to be applied is an impressive statement that the journey has started.

More emphasis on the case studies from those areas of research and less on poorly substantiated benefit claims would be appreciated.