Very good talk Ruairi.

I think you and the gentlemen from the audiance are actually approaching the same subject from different angles. You’re growing a complex behaviour in a petri dish and he’s taking one apart to see how it works.

I understand where you’re coming from on the lack of understanding of the brain, A lot of work is theory, and they remain theoretical until they can be tested

But we don’t need to have a complete understanding of the brain before the information of what we do know becomes relevant. The moon landings were done with what we now consider to be very simple science.

Somewhere between an interactive environment and ‘us’ is our brains so it definetly plays a part at some point. Interacting with something is a dialogue, it is a two way event, or a communication. We can only have a dialogue with something through some kind of language; moving your body through the room, facial expressions, making sounds or whatever else it may be. There is a transition of information between someone and the thing they are interacting with. Some kind of bridge between the mental internal world of experience and the reality that actually surrounds us. It’s like a land bridge between two subject areas!

We may not have succeeded in getting AI-like results from remodelling our understanding of the brain. But we can still produce highly engaging and meaningful interactive surroundings by applying our models of the brain to them. As a survival mechanism in intelligent forms of life, most animals have intention. Normally this is, procreating, or eating. Intention is in-built, or is learnt and is normally enforced by reward for doing well and punishment for doing something wrong. For a behaviour that we would consider meaningful to arise we could do well to begin with this as a basis for the running program.

Hi David

Yes your absolutely right. When I said no, this is actually all my work, I guess I should have mentioned Charles Babbage, Herman Hollerith, John Von Neumann… ok I’m being a smart ass and I see your point, that of course I was working on the shoulders of giants, but I was being genuine in my answer. I have tackled all the design challenges myself both design and construction as well as behaviours and learning algorithms.

So to credit some of the platforms on which this was made. I used the OpenCV library for Facial recognition. The Genetic Algorithm was based on looking at some flow diagrams given to me by Sean Hanna, a research at the Bartlett who I wrote about recently on my blog. The GA was implemented in Processing and the OpenCV library was implemented in MaxMSP using jit.CV. The robotics were controlled by Arduino I/O boards. Regarding papers, if anything I think Gordon Pask’s ‘A Comment, A Case History and a Plan’ was the only piece of text that really had a major impact on this installation and that was not in a technical sense. Perhaps I would also like to credit Rodney Brooks’ work on Mobile Robotics as an inspiration.