A Framework for Scientific Discovery through Video Games

As science becomes increasingly computational, the limits of what is

computationally tractable become a barrier to scientific progress. Many

scientific problems, however, are amenable to human problem solving skills

that complement computational power. By leveraging these skills on a larger

scale---beyond the relatively few individuals currently engaged in

scientific inquiry---there is the potential for new scientific discoveries.

This book presents a framework for mapping open scientific problems into

video games. The game framework combines computational power with human

problem solving and creativity to work toward solving scientific problems

that neither computers nor humans could previously solve alone. To maximize

the potential contributors to scientific discovery, the framework designs a

game to be played by people with no formal scientific background and

incentivizes long-term engagement with a myriad of collaborative or

competitive reward structures. The framework allows for the continual

coevolution of the players and the game to each other: as players gain

expertise through gameplay, the game changes to become a better tool.

The framework is validated by being applied to proteomics problems with the

video game Foldit. Foldit players have contributed to novel discoveries in

protein structure prediction, protein design, and protein structure

refinement algorithms. The coevolution of human problem solving and computer

tools in an incentivized game framework is an exciting new scientific

pathway that can lead to discoveries currently unreachable by other methods.

© 2014 ACM Books (E-bok): 9781627055062

Utgivningsdatum

E-bok: 1 juli 2014