In <i>Bring Home Water</i>, players must move one valleylike potential (left) to capture the quantum wave in another (right) and then bring the wave back to a drop-off zone.

In Bring Home Water, players must move one valleylike potential (left) to capture the quantum wave in another (right) and then bring the wave back to a drop-off zone.


Humans best computers in atom-snatching game

In the latest effort to entice ordinary people to do scientific scut work, physicists have enlisted online gamers to figure out the fastest way to pick up and move an atom with a beam of light. Surprisingly, people playing an online game came up with better strategies for moving the atom than a computer algorithm alone. Indeed, they found solutions that were faster than what the physicists had assumed was a speed limit set by quantum mechanics itself.

"What we thought was the quantum speed limit based on the numerical simulations was not, in fact, the quantum speed limit," says Sabrina Maniscalco, a theoretical physicist at the University of Turku in Finland who was not involved in the work. Tommaso Calarco, a theorist at the University of Ulm in Germany who also was not involved in the work, notes that the new game differs from most crowdsourced science games, which rely mainly on humans' ability to recognize images and spatial patterns. "It's a different kind of visual intuition," he says. "What I love about it is that they don't know how the players did it."

Known as Bring Home Water, the game is part of a suite of games called Quantum Moves, the latest in a growing list of crowdsourced scientific games. Since 2007, "citizen scientists" have helped astronomers classify millions of galaxies in the Galaxy Zoo project. For nearly as long, they've helped structural biologists figure out how proteins fold in the game Foldit. And since 2012, they have helped neuroscientists trace the spidery neurons of the brain in the game Eyewire. Those games depend primarily on a person's ability to recognize spatial patterns—for example, you know an elliptical galaxy when you see one.

In contrast, Bring Home Water relies on people's knack for performing tasks that involve dynamic movement. Users must pick up and move an atom with a simulated spot of laser light. To be accurate, the atom and the laser beam must be modeled according to quantum mechanics. So in the game the atom is not a simple ball, but is represented by an extended, rippling quantum wave. The height of the wave at any point gives the probability of finding the atom at that location. To begin, the atom's quantum wave fills a valleylike "potential," which serves as a trap that holds the wavelike atom in its initial position. The laser beam is a second valley that user can move sideways and make deeper or shallower. The trick is to get the quantum wave to “slosh” from the fixed potential into the movable one and then to cart it back to a drop-off zone. Such manipulations are routinely done with real atoms trapped in spots of laser light and tugged about with "optical tweezers."

Jacob Sherson, the physicist at Aarhus University in Denmark who led the team that developed the game, says he expected people would fail miserably at the task. "I thought we would do this, and we would find out that it doesn't work," he says. "That's been one of the biggest surprises, that if you give people 1 or 2 seconds they come up with solutions that are better than any that a computer comes up with."

Sherson and colleagues compared the performance of people with those of a computer program known as the Krotov algorithm, which started with random "seed" trajectories and then searched for the fastest way to retrieve the atom. If given enough time, the computer program snatches the atom with complete reliability. Below a certain time, however, the program's success rate plummets to zero. Physicists had assumed that the falloff resulted from a fundamental quantum speed limit set by parameters such as the maximum depth of the potential representing the laser beam.

However, human users were able to find ways to move the laser beam that worked faster than the computer algorithm alone, Sherson and colleagues report online today in Nature. When those human solutions were used as starting points for the computer algorithm, it worked 30% faster than on its own. Moreover, people found two general classes of movements. In the first, they moved the valley for the laser right next to the one for the trap to produce a double valley and waited for the wave to slosh into the other. That "tunneling" solution is what a physicist would try first, Sherson says. In the second, they moved the laser valley past the trap and made it deeper, causing the wave to rebound back toward the drop-off region. That novel shoveling solution shows the potential for gamers to discover new things, Sherson says.

Bring Home Water has about 10,000 players and is probably too simple to be wildly popular, Maniscalco says: "It's still not one of those games that people play like mad like Angry Birds." Still, she says she expects to see more games involving so-called quantum control. In the meantime, Bring Home Water has left quantum physicists with a puzzle, Calarco says. They had assumed that computer algorithms such as the Krotove algorithm conk because they've hit the quantum speed limit. Now, it appears that those analyses were mistaken. Physicists plan to hold a conference this June, Calarco says, to figure out how they can determine what the quantum speed limit really is.