Toby Simpson is Chief Technology Officer at Ososim. He spoke to a group of journalists for London Press Club, all of whom were concerned about their jobs. He explained the the possibilities of Artificial Intelligence (AI) stating that it will be like having a whole bunch of researchers working for you, enabling you to go further and wider.
His field is the world of machine based learning; he firmly believes in providing software solutions where complexity emerges from the interactions of simple components. Learning brought to life. Below you can read his comments from the event.
Artificial Intelligence can play games
Just because AI can play GO or Chess, that is one very limited possibility. It is a bit like saying if I jump high enough I can reach the moon, and each time you say I’ve jumped higher than before, but you’re still a long way away.
Being able to isolate, analyse and plan are attributes that AI hasn’t yet begun to approach.
The mind is a huge simulation machine which disappears into many many parallel alternate universes. As you explore in your mind all these weird journeys, you weigh them all up and a lot of them you’re never going to do. All going on at the same time, constantly running a huge variety of possibilities. Human beings can’t just think about one thing – there is constant analysis, planning and experimentation going on.
Take a leaf out of nature’s book. Nature does things with large populations of simple things which combine, out of which this amazing complexity emerges. That’s what we mean by biologically inspired intelligence, keeping it simple.
Ososim is trying to simplify things particularly from a learning perspective. We like that approach of managing complexity. We like that biological inspiration. It’s not just for the intelligence, it’s the way you structure your software, to create your simulations.
With AI in educational terms, eventually you will be able to go anywhere to any place at any time, in complete safety. Take your class to stand on the edge of Vesuvius, or travel through a jungle without fear of the animals or insects.
It’s a philosophical approach to development. You don’t have to understand the problem to be able to solve it. You don’t even have to know how you got there. If you try to emulate rather than model, you make a great list of possibilities and even then you never cover all the options.
But if you put all the right building blocks in place, and let the simulation unfold you will end up with the right solution. If your model is good enough, what will come out of it, will be plausible and it won’t be a syntax error. It will all make sense even if you did not specify it or expect it.
If you keep it simple and don’t over complicate it, then it’s easy to achieve one of the learning goals. If you had to design, into modern flight simulations, every condition under which a plane could stall you would end up with a thousand different conditions and you wouldn’t have even started. Furthermore, a real pilot would be able to tell you immediately that you’ve got one wrong.
Instead you model flight dynamics, you model the lift dynamics, you model the process. You put all those things in place and then when the plane stalls, it stalls for all the same reasons as it would in reality. You’ve built a model of the environment and it behaves exactly like a real plane, without anyone getting hurt.
When it comes to creating very large scale simulations they must be believable ones as well because if you’re putting someone in a training environment, say for stake holder management or a merger, you want people to put aside their disbelief and be involved in the simulation. Some of the best learning, or education, is when you don’t realise you’re learning.
Learning is a state of mind, you have to want to learn. Curiosity is a wonderful gift to have. You give that to a young person and they’re armed for life. The greatest weapon we’ve got for our future is quality education. Unfortunately now we live in a world increasingly where people are proud to be ignorant. I don’t think ignorance is anything to be particularly proud of. Once you accept that it’s cool to be ignorant, you immediately throw away the desire, the requirement to learn anything new.
SIRI and the importance of context
One of the things that human beings are very good at, is relevance and isolating.
Take Siri. If take out your smart phone in the middle of a busy party and ask Siri something it will be laughably bad because it will pick up all the conversation. However, as a human you are able to isolate the one conversation in all those loud voices. Even more importantly, although you only hear 20% or 30% words that are said you are able to piece together the missing words. That’s why Siri and all these other clever things are so far from real intelligence, but they are useful AI augments rather than supplant. It will be intelligent but in a different way. It will be different because a human brain has billions of years of evolution behind it. I guess none of these pieces of tech are a bad thing, but it makes the world a smaller place.
Girls into STEM
Any problem that you want to solve is always made better with diverse minds and you will get a better result. Else you have a room of alpha males trying to alpha male each other.
I have two daughters and I wouldn’t want them to believe that opportunities are not available just because they are female.
For Toby’s Advice please see the last post.