Excerpt: How Turning Players Into Data Processors Is Changing The Game
In an exclusive pair of excerpts from David Edery and Ethan Mollick's book 'Changing The Game: How Gaming Is Transforming The Future Of Business'
, the duo examine how web-based games such as The ESP Game and Fold.it can be used to get humans to process important data sets -- for free!
The book, which has an official website
with more information about it, discusses "...how leading-edge organizations are using video games to reach new customers more cost-effectively; to build brands; to recruit, develop, and retain great employees; to drive more effective experimentation and innovation; to supercharge productivity... in short, to make it fun to do business."
In addition, its introduction notes of the wider field of games used for non-entertainment purposes:
"Companies of all shapes and sizes have begun to use games to revolutionize the way they interact with customers and employees, becoming more competitive and more profitable as a result. Microsoft has used games to painlessly and cost-effectively quadruple voluntary employee participation in important tasks.
Medical schools have used game-like simulators to train surgeons, reducing their error rate in practice by a factor of six. A recruiting game developed by the U.S. Army, for just 0.25% of the Army’s total advertising budget, has had more impact on new recruits than all other forms of Army advertising combined.
And Google is using video games to turn its visitors into a giant, voluntary labor force -- encouraging them to manually label the millions of images found on the Web that Google’s computers cannot identify on their own."
The book's authors, which include Gamasutra contributor and Microsoft Xbox Live Arcade staffer Edery, particularly recommended these intriguing sections of their new title, dealing with the last of the three examples mentioned above:
The ESP Game
It turns out that there are many areas where human intelligence is still superior to computing power.
For example, computers are very bad at identifying images, which has become an increasingly important problem as millions of new photos and illustrations are uploaded to the World Wide Web each year.
Luis von Ahn, Assistant Professor at Carnegie Mellon University and winner of a MacArthur Foundation "genius" grant, developed The ESP Game
to help solve this problem.
In The ESP Game, two anonymous players are matched online without any means of communicating with one another. Both players are shown a random image (for example, a flowering plant under a clear sky) while a clock counts down two and a half minutes.
The players must then type words that describe the image, such as "plant," "flower," "pretty," or "sky." When both players have typed at least one word In common, like "sky," they both score points and a new picture is shown. At this moment, the players have helped teach the computer that this picture contains a "sky."
This goes on until the players have run out of time. In an interesting twist, sometimes pictures get recycled by the game after they have already been labeled once.
When this happens, the old label (i.e. "sky") is no longer accepted by the game, so players must come up with a second or third word to describe the picture. In this way, the game forces players to give each picture a more detailed description.
The ESP Game is undeniably addictive. Many people play for over 20 hours a week, and over 20 million labels have been harvested in just a few years; the equivalent of several million dollars of free labor.
Professor von Ahn estimates that just 5,000 people playing The ESP Game for a month - a tiny number, compared to the active populations of many gaming websites -could label every image on the Web.
In fact, Google found the ESP Game to be so useful that it licensed the game for its own use, as the "Google Image Labeler
The ESP Game is just one of a variety of "games with a purpose" developed by Professor von Ahn. Peekaboom refines the data gathered by The ESP Game by encouraging players to identify the parts of a picture that are associated with a given label.
For example, two players of Peekaboom might be shown the picture of the flowering plant under a clear sky. The first player, seeing the label "sky", might correctly click the blue region above the flowing plant.
The second player must then guess the correct label, solely by seeing what region of the picture was selected by the first player. If the second player guesses "sky," both players score points.
The techniques pioneered by Professor von Ahn may soon be applied to many other real life problems. For example, the Transportation Security Administration has expressed interest in creating a game where "bags are screened at the airport and sent over secure networks to multiple players who help determine the contents of each image.
This could imply a major gain in security as an aid in the baggage screening process: instead of a single officer looking at each bag, multiple people could see each bag, having a higher chance of finding potentially dangerous objects.
Fold.it And Distributed Innovation
Distributed innovation games take a different approach to gathering information than human computation games.
Whereas human computation is powered by large groups of people playing a game as frequently as possible, innovation games are designed to identify and bring together those rare individuals who can creatively solve very difficult problems. In other words, human computation games are about manpower; innovation games are about brainpower.
is an example of an innovation game - one that might soon help develop new treatments for disease by exploring the secrets of protein folding.
Proteins are the molecular structures that drive most of the important functions in living beings; their shape determines how they interact with other substances. A particular shape might make a protein effective at clotting blood, while a differently-configured protein might prove effective at neutralizing the HIV virus.
Unfortunately, determining the ideal shape of a protein can be very difficult. Even a small protein with less than 100 chemical components could have a huge number of configurations - about 3^100, to be precise.
There are algorithms that can be used to determine whether one protein shape is better than another, but taking advantage of them requires computers to test billions of possible shapes for each protein, which requires massive amounts of computing power. So Fold.it takes a different approach to the problem, capitalizing on people's spatial awareness and problem-solving skills to achieve what computers cannot.
Fold.it turns protein folding into a game, where the goal is to score the greatest number of points by identifying the optimal shape of a protein. Using a colorful graphic interface, players push and pull pieces of a protein into various shapes, trying to maximize their score by eliminating chemical incompatibilities.
A team-based component of Fold.it ensures that players will work together to solve particularly tough problems. And Fold.it's designers can create protein puzzles for any purpose, such as the best configuration to treat a particular cancer, and then have the players try to solve it.
One of the fascinating discoveries to come out of Fold.it is that the best solvers of a particular problem may not be the people you'd expect. In the words of Fold.it designer Seth Cooper, "Some of the top scoring players are biologists, but the people who are really doing well and consistently winning don't have any biological, or even academic, background at all."
Unlike a human computation game, distributed innovation depends on finding the best people to solve a particular problem, no matter who they might be. Indeed, most players of Fold.it give up quickly and move on, but those who stay tend to be quite talented, and many become addicted to the game. And the Fold.it team has every intention of leveraging the efforts of those talented players for the benefit of humankind.
They plan to create games that will help design biofuels and vaccines, and express the hope that someday, an innovation game player will win a Nobel Prize for the work they did while "playing around."