Red King and crowd computing

Posted April 6, 2016 by Amber Griffiths

We've started a new project called the Red King, in collaboration with Ben Ashby. The researchers have developed models of host and parasite evolution, and are interested in what conditions need to be met for diversity in host and pathogen types to arise. Originally we were asked to make an educational game for outreach purposes – but it's much more interesting if we can make this a two-way exchange. We're looking at whether we can use crowd computing to turn this into a citizen science project.

Crowd, volunteer, or distributed computing is contributing your computer's idle processing time to contribute a small part to a large centralised project. This approach is well established in scientific research, and most projects use the BOINC platform (Berkeley Open Infrastructure for Network Computing) which was created in 2002 by the University of California, Berkeley.

BOINC was originally developed for the SETI@home project, looking for extra terrestrial communications from radio telescope data. Since then it's been used a large number of projects spanning mathematics, linguistics, medicine, molecular biology, climatology, environmental science, and astrophysics.

I'd assumed it would be painful to set up, even as a user, but it really is very straightforward. I set up BOINC and started climate simulations going in under five minutes. The download is available through the Ubuntu Software Centre, then you just pick the projects you want to contribute to. It looks like the process is similarly easy on Mac, Android and Windows. As I write this it is ticking away in the background, having no noticeable effect on the performance of the computer.

It's not pretty, and the only reward is a warm fuzzy feeling, but it is seems to work perfectly and there is essentially no barrier to getting started. If the installation process had been cumbersome, I might have moved on. I'd like to see more about what my computer is doing – some visualisation or explanation of the specific model it is working on. Being able to access the data I create would be good too. Agreeing to participate in crowd computing requires trust – not only trust that nothing nasty is going to be put on your computer (as for any software download), but trust that your computing power won't be used for anything other than what you've agreed to. This might be a major barrier to it becoming more widely used.

I'm hoping we can make a more bespoke system, with the users making the research decisions. At the moment we're thinking about asking people to set two lots of model parameters, and have their computer generate every model in between. We'll the sonify this data, and ask users to tag places where the sound changes, indicating shifts in the model outputs. They can then use the sonifications to make music.