Friday, May 16, 2008

Problem Solving With Video Games

There is a Video Interview With Peter Molyneux where he talks about not being able to do much else but design games, and he wonders if he could apply his knowledge of gaming to other human pursuits and endeavours.

I thought this was an interesting idea, then I realised that I had meant to post about a project that does just that. So here we are.

There are some problem domains in the world where to solve a problem or improve something you have to understand a process completely. Programmers of Artificial Intelligences have been known to spend exorbitant amounts of time looking at quite simple issues so that they can create AIs that model intelligent behaviour. Good examples of this are anything that uses Fuzzy Logic. That is anything from Rice Cookers to Lifts, Thermostats and a huge amount of Control Systems.

There are some problems that are so difficult to understand that they can't yet be translated into software, because we don't fully understand what is going on. The problem I am referring to here is Protein Folding. We know what the acid code of the protein is, we know what it looks like in 3D when it has finished folding, but we don't know how it folds into it's final shape or how it moves, so we don't know what it does.

This is where Fold It! from the guys working on Rosetta@home comes in. It is a game where you as the player are tasked with folding a protein. Higher points are rewarded by finding a structure that requires the lest energy for the protein to fold into. The idea of this is that players will eventually discover general tactics that can be worked into protein folding algorithms and a rosetta method of protein folding be found. From this new drugs could be created in the lab fit for purpose artificially and our understanding of our own bodily processes greatly increases. We might even be able to find a cure for ageing.