Machine learning AI has landed
The concept of Artificial Intelligence (AI) has been a staple of science fiction and futurism for over 50 years. But with recent developments, it is poised to make an increasingly real difference to our future.
To talk us through it, our Global CTO Dr Joseph Reger wowed a packed auditorium in his AI keynote at Fujitsu Forum 2016.
It’s a complex and fascinating subject, so in this post I wanted to bring you the highlights from his talk…
A bit of AI history
The story of AI started back in 1956 at Dartmouth College in the United States. Initially scientists proposed a two-month study. Needless to say, the challenge wasn’t cracked quite so quickly.
The term for AI was coined, however – a brand of computer science attempting to build machines capable of intelligent behaviour.
Now in theory you could build a machine able to look at all inputs and eventualities of a situation.
“But that wouldn’t be very intelligent,” said Dr Reger.
This is where machine learning comes in. Dr Reger explained that by definition this is the ability for a machine to learn instead of being explicitly programmed.
These systems can look at “many, many examples” of information, he said. We might then give some help and feedback on how they ‘learn’.
He explained that in a machine learning system it’s the quality of the data being fed in that’s important.
What advances have we made?
By applying machine learning with neural networks, computer networks modelled on the human brain, you can massively increase the bandwidth available for processing data.
In doing this, a machine can determine patterns among images and pictures much more easily than other data sets.
This means you can apply AI in some very interesting ways.
You can make it work with existing infrastructure. So if you have a CCTV camera monitoring the numbers of spaces in a car park, for instance, all the AI needs to see is the feed from the camera – saving costly investment in sensors and infrastructure.
Fujitsu already has a working use case for this!
— Fujitsu UK (@fujitsu_uk) November 16, 2016
This means in the world of the booming internet of things, AI can have a direct impact on how we collect and make sense of information globally.
By taking existing data sets and turning them into images, an AI can process vast amounts of information very quickly, enabling better human decision making.
AI securing our future
If you consider the huge amount of information that passes through the internet at any one point in time, no human alone is going to be able to detect every hidden pattern at once.
But AI could be used to spot anomalies in that traffic at scale and speed, meaning the time it takes to detect cyber threats is vastly reduced.
Overall, Dr Reger’s talk left no doubt that adoption of AI will grow rapidly over future months and years – and will be a huge growth area for our industry.
This article obviously barely scratches the surface of everything covered at Fujitsu Forum – if you’re interested, I highly recommend watching the full keynote from Dr Reger himself below:
Latest posts by Jim Millen (see all)
- How education and industry are disrupting as one - July 13, 2017
- Fujitsu World Tour 2017: 8 things we learned about disruption and co-creation - July 6, 2017
- 5 fascinating questions from this morning at Fujitsu World Tour - July 6, 2017