FoldIt and "Citizen Science"

-- By AaronChan - 19 Oct 2011

Online Models of Volunteerism

Information access does not need to be unidirectional. Instead of traditional broadcast models, information is a two-way street on the Internet. Online access is a tool for empowerment, not just by giving power to users, but also in allowing them to contribute. The Web is now a vast depository of human knowledge; it can be leveraged for greater societal goals. For example, the success of Wikipedia demonstrates what can be accomplished with collaborative experience.

Michael Goldhaber suggests that motivation on the Internet is driven by reputation. Yet whatever the form the remuneration takes, it is still premised on a quid pro quo. Yet reputation would lose its incentivizing effect as the activity becomes increasingly dispersed and individual contributions are further anonymized. This is precisely the environment necessary to harness the widest participation. I focus instead on the frequently overlooked incentives of satisfying human curiosity, problem solving, and amusement as driving forces for scientific research.

These incentives drive the human attraction to games. Harnessing these incentives joins gaming with research in a way that channels mindless diversion into productive causes.

Puzzles and the Role of the Integrator

FoldIt is computer game designed by University of Washington researchers that allows players to design protein structures by folding amino acids. Given the complexity of folding dozens, if not hundreds, of amino acids into viable protein structures, brute force computing power would not be a practical way of solving the problem with the current technology. Humans have an advantage in recognizing patterns and certain three-dimensional spatial intuitions that computers cannot currently simulate. What had stumped scientists for more than a decade, a retrovirus protease in rhesus monkeys, was solved by a worldwide consortium of players in ten days. The vast majority of these players had little biology or chemistry experience, relying only on the rules of the game as dictated by real physics. The discovery of the retroviral protease has been published in a scientific journal, naming a team of players, the FoldIt Contenders Group (FCG), as coauthors. In an interview with one of the FCG members, “mimi” preferred to be credited pseudonymously or anonymously. The main motivation for many of these players was advancement of science, not fame or recognition. They were not paid for their participation; many simply wanted a challenge and to work towards something big.

Video games, at their most fundamental level, are sets of puzzles. Rather than create arbitrary puzzles, game developers can insert real life puzzles within games. For example, reCAPTCHA’s verification system uses human activity that would otherwise be wasted to digitize books. Human effort is already being expended; it can be channeled into more productive uses as long as it is made sufficiently entertaining. Those in need of mass problem solving in discrete, miniature parts can outsource their puzzles to game developers.

FoldIt is only one of several other projects collectively grouped together as “citizen science.” These projects crowdsource because they benefit from certain human attributes that are not duplicable in computers currently.

Yochai Benkler's paper Coase’s Penguin, or, Linux and the Nature of the Firm evaluates the motivations for contribution to the open software movement. He identifies the spontaneous creativity birthed when components from production are “sufficiently small grained, and the cost of connecting people to projects sufficiently low thanks to cheap network connections.” By breaking difficult project into discrete consumable parts and make participation as least cumbersome as possible, crowdsourced projects can harness the creative spark in a wide participatory network. While this phenomenon has been documented before, I suggest that there is an additional incentive that is a driving factor for FoldIt—specifically, gaming taps into satisfying human curiosity, problem solving, and diversion. Humans naturally seek patterns and order. Puzzle solving is core to this inner desire.

Tapping into the non-gaming online behavior has been a critical model for the success of such projects as Google, which reinforces its search algorithms based on the conduct of its users. People may play games because they need diversion or because they intuitively seek reward and games feed on that hunger for instant gratification. Whatever the psychological motivations behind gaming, that activity can be channeled. Given the tremendous amount of energy devoted to quick and simple online games (Farmville at its peak had 83 million monthly active users) some of that mindless clicking can be channeled into projects with more social impact beyond increasing sales for game companies.

It takes what Benkler calls an “integrator” to manage those clickstreams and combine the discrete component functions. He identifies four mechanisms for solving the integration problem.

1. Iterative peer production of the integration function itself
2. Technical solutions embedded in the collaboration platform
3. Norm-based social organization
4. Limited reintroduction of hierarchy or market forces to provide the integration function alone.

For scientific research, the first mechanism may not be appropriate unless the players are all sophisticated because it would require a high level integrator to atomize the tasks and divine larger meaning from their outcomes. Game designers can solve the integration problems either embedding technical solutions within the puzzle platforms, creating a social organization that encourages integration, or by introducing a hierarchy to the puzzles. In other words, scientists can outsource the integration function to designers.

Scientists can team up with game developers to create a commercially viable video game with puzzles that have real-life applications. Only a well-designed game can attract the type of click workers needed en masse for some research activities. Developers are not looking to control the outcome of their puzzles; they generally only seek to provide the highest level of profitability. With the exception of some developers who work with an eye towards aesthetic, most developers care most about a stimulating experience. In video games, that often means the highest level of amusement since the value of video games lies almost entirely in their value as diversion. While not all tasks are easily commercialized in this way, scientists should consider gaming as a source of labor.

Scientists may benefit from game designers who can help craft seemingly mind-numbing research tasks into appealing challenges and puzzles. While folding a protein to achieve a particular shape in itself may not be appealing to the gamer population, the FoldIt? designers provided ease and competition for the activity. They are game designers in so much that they dress the actual scientific problems to make them accessible to those with no scientific background. The game is not in the puzzle itself, but the puzzle is only one aspect of the game. Instead of creating puzzles, designers can repackage actual problems and integrate them into the game. With the iterative feedback of web-connected games, these puzzles can be continually tweaked, replaced, or improved upon.

With the evolution of online, collaborative gaming, even if there are not deliberately inserted scientific puzzles, the scientific community can benefit from observing gaming sociology and group dynamics. Economists can apply their theories to in-game trade economies. Sociologists can examine avatar gender preferences. The passive activity generated from participating in games can be tapped either through intentionally planted puzzles or from merely observing group behavior.