backtop


Print 30 comment(s) - last by Camikazi.. on Sep 20 at 2:59 PM


  (Source: South Park Studios/Comedy Central)
It's the first time gamers have found the solution to a longstanding scientific problem

Video games may not enhance cognitive abilities, as concluded by a recent Florida State University study, but that doesn't mean that gamers aren't capable of achievement beyond just saving the princess.

Gamers have successfully figured out the structure of an enzyme of a virus similar to AIDS, which has baffled scientists for about a decade.

Scientists at the University of Washington have been stumped by the structure of a monomeric protease enzyme, which is an agent in the "complex molecular tailoring" of the family of retroviruses that includes HIV. Scientists could only see a one-dimensional, flat image of the protein through microscopes and needed a 3D image in order to unravel the molecule. By doing this, researchers could reveal potential targets for drug treatments and better understand how certain diseases work in general.

In 2008, the University of Washington researchers developed a video game called Foldit, which challenges teams of gamers to unfold chains of amino acids (the building blocks of proteins) using certain tools.

Gamers were divided into competing groups, and played the fun-for-purpose video game. In just three weeks, the gamers had created an accurate model of the enzyme. According to Yahoo News, it's the first time gamers have found the solution to a longstanding scientific problem.

"We wanted to see if human intuition could succeed where automated methods had failed," said Firas Khatib, of the University of Washington's biochemistry lab. "The ingenuity of game players
is a formidable force that, if properly directed, can be used to solve a wide range of scientific problems."

Seth Cooper, one of Foldit's creators, explained that the gamer's succeeded where the computer failed because people have spatial reasoning skills, and the game further brought the strengths of humans as well as computers together.

Scientists are now using the knowledge of the structure to try to develop better treatments for retroviruses like HIV.


Comments     Threshold


This article is over a month old, voting and posting comments is disabled

hmmm
By Joz on 9/19/2011 4:42:09 PM , Rating: 5
That is impressive (IMO). Now, get to work unraveling the mysteries of warp speed and matter to energy to matter transportation over large distances.




RE: hmmm
By Shig on 9/19/2011 6:20:18 PM , Rating: 2
Most distributed computing projects are transforming their respective fields. Folding @ Home is another example.


RE: hmmm
By BugblatterIII on 9/19/2011 6:49:14 PM , Rating: 3
Yes, but this was using people's brains rather than their PC's computing power.


RE: hmmm
By cmdrdredd on 9/19/11, Rating: -1
RE: hmmm
By Camikazi on 9/19/2011 9:00:16 PM , Rating: 4
ou In 2008, the University of Washington researchers developed a video game called Foldit, which challenges teams of gamers to unfold chains of amino acids (the building blocks of proteins) using certain tools.

Gamers were divided into competing groups, and played the fun-for-purpose video game. In just three weeks, the gamers had created an accurate model of the enzyme. According to Yahoo News, it's the first time gamers have found the solution to a longstanding scientific problem.

Did you miss that part?


RE: hmmm
By Dr of crap on 9/20/11, Rating: 0
RE: hmmm
By geddarkstorm on 9/20/2011 12:08:18 PM , Rating: 2
The computer didn't do the computations, it just gives the interface. It's the human brain that did the solving--that's the whole idea behind Foldit. The Foldit program is very amusing, if you've ever tried it; and this is definitely an amazing result.


RE: hmmm
By Camikazi on 9/20/2011 2:58:52 PM , Rating: 2
The game just gave the tools, the people and their brain power actually did the work. That was the point of the game, computers could not figure this out but people using the same set of tools did, the gamers did it not the computer.


RE: hmmm
By FITCamaro on 9/19/11, Rating: -1
RE: hmmm
By inaphasia on 9/20/2011 7:12:43 AM , Rating: 2
Careful what you wish for. You'll probably end up standing in queues.


RE: hmmm
By Goty on 9/19/2011 10:27:01 PM , Rating: 1
There's no mystery to warp speed, it's just not technologically feasible. There's a nice little section in Hartle's "Introduction to General Relativity" that lays out the physics quite clearly.


RE: hmmm
By Bad-Karma on 9/20/2011 12:41:33 AM , Rating: 2
Where General Relativity fails in the concept of "light/warp" speed, Quantum Mechanics picks up and says "sure, why not?" There are several scientist working in the Quantum field with plausible theories of faster than light travel.


RE: hmmm
By jecs on 9/20/2011 2:42:39 PM , Rating: 2
Yes some scientist believes this could be achieved by contracting the space in front of the spaceship and expanding the space behind. There is not a law in physics preventing the space to contract or expand faster than light. Not yet, that is.


RE: hmmm
By jeff834 on 9/20/2011 3:35:44 AM , Rating: 2
Are you saying it is not possible to travel faster than light or to create a "bubble" and bend space around it? I would agree that one could not accelerate to the speed of light, however I don't know of any undeniable proof that the latter would not be possible. Why could we not one day create stable Enstein-Rosen bridges? I mean not tomorrow, but one day.


RE: hmmm
By Shadowmaster625 on 9/20/2011 11:11:38 AM , Rating: 3
"The Vulcan Science Directorate has determined that time travel is impossible"


RE: hmmm
By Camikazi on 9/20/2011 2:59:17 PM , Rating: 3
Never trust those sneaky Vulcans.


HIV
By Salisme on 9/19/2011 4:46:01 PM , Rating: 5
quote:
Gamers have successfully figured out the structure of an enzyme of a virus similar to AIDS


AIDS is not a virus, it is a condition caused by a virus.




RE: HIV
By inperfectdarkness on 9/19/2011 7:12:22 PM , Rating: 2
agreed. HIV. big difference between syndrome and virus.


RE: HIV
By Breathless on 9/20/2011 9:28:09 AM , Rating: 2
Bingo


RE: HIV
By geddarkstorm on 9/20/2011 12:09:25 PM , Rating: 3
Shhh. Quiet, you and your facts.


Which was first...
By BugblatterIII on 9/19/2011 5:00:39 PM , Rating: 5
...this or Stargate Universe?




RE: Which was first...
By Joz on 9/19/2011 6:59:48 PM , Rating: 2
lol, I'd + that, but I can't :(

+6 it.


Read the link for a better understanding
By Hieyeck on 9/20/2011 10:43:21 AM , Rating: 4
tl;dr:
Computers can do better calculation, Humans can do better reasoning. 1+1 = The whole is greater than the sum of it's parts. Humans can filter out the 99.99% of the wrong guesses, so computers have 10000x less probabilities to run through and produce results 10000x faster.

Practical Demonstration:
http://www.bridgebuilder-game.com/bbg-info.php

A computer wouldn't know what a bridge is and you're trying to tell it to build a bridge? You would know that it would have to cross the span. A computer might eventually get the right shape, but stick it 30 miles in the air. All it would know is that it's a wrong result and would start with a jumble of sticks again. On the other hand you could build a bridge, let the computer test to see if it will stand, know where it's weak, where you can save weight and money, and make a result that's more successful, eventually creating a working bridge that would not only carry itself, but a train across.

Layman's (AKA I'm at work and can't download BBG) Explanation:
Nature has a tendency to work towards efficiency (that's why bubbles are spheres, evolution works, etc.). The 'game' is a computer program which can calculate the total energy used to construct the molecule. The game would award more points for using less energy. Scientists could check the constructs that provided the highest scores to verify.

So why can't a computer just construct it in the first place? The problem is that the computer has no sense of intuition. Imagine a dial, say at the 3 o'clock position, and you were told to get it point to 12 o'clock. Humans would intuitively know to turn it counter-clockwise. Computers, on the other hand, don't have a sense of clockwise or counter-clockwise, and would have to try every single position - it might even get it right on the first try, but it would need to compare the result to all its other results to realize the first one was right. Now imagine you're not using clock hours, but degrees in a circle. That's 30 times more results. Then you can get into minutes and seconds of degress and suddenly it's starting to look sketchy. AND you're rotating just ONE dial. This protein has hundreds of dials. And once you rotate one dial one unit, all the other dials reset and you have to start from scratch to generate the results again. AND you're still only changing the ONE dimension. Imagine, having to generate results for not only the rotation (rotation on Length), but the flip of the dial (rotation on Width). And then add on that the location of the dial on a grid (rotation of Depth), since this is a 3D world. Oh nuts.

(excuse the poor grammar, I'm using it illustrate human vs cpu) So where do humans come in? As I've said, Humans would just know to turn that one dial counter-clockwise because they've built on previous experience that counter-clockwise is the shortest direction. On top of that, humans have a sense of "-ish", or approximation. So even if all the dials reset, we can turn each dial back to 12 on our first attempt and KNOW it's right (or in this case, it can be turned back to 12-ish and know it's right-ish), cutting back on the number of attempts, and allowing us to move on to the next dial sooner.

Still, with a supercomputer with millions of cpus and billions of threads working in parallel, it can generate a HUGE number of results and compare them. However, 99.99% of those results are wrong-er and this highlights another advantage humans have. We don't need 99.99% of the wrong-er results to know the 0.01% right-er-ish results and take a stab at it. Simply put, we can get to right-er-ish needing billions of times less guesses, because we can make educated guesses. On top of that, humans can share. It's not ONE human, it's THOUSANDS of us making right-ish guesses and each of us can only get right-er as we share that information. Each human processing 'thread' learns from the other thousands of 'threads' in real-time (via shared high-scores) and better results are generated exponentially faster.

The human weakness is that it's difficult for us to determine which is the best result, because of the intense calculation required to determine the energy use in the protein (or back to the dial analogy, we can't turn them very accurately). That's where us working with computers made this such a success. The ability of the computer to make the quick calculations allows us to learn what's right-er and wrong-er as we keep working at it. If we overshoot turning a dial, a machine can tell us that the result previous was a better and we can tap the dial back into place. Over hundreds of dials, we can easily set them to 12-ish and THEN tell the computers to only go between 11:59 and 12:01 and let the computer figure out how many seconds and milliseconds we were off.




By geddarkstorm on 9/20/2011 12:17:06 PM , Rating: 3
Every neuron in your brain by itself can do pattern recognition. Put the whole brain together (over 100 billion neurons), and you can do calculations no computer can (even the math required is often too sophisticated to be realistic). There are things computers are much better at like straight up computation, but pattern recognition is where they fall short and we excel. (pattern recognition is what's important in nature, not computation)

When it comes to folding proteins, pure computations are things like Molecular Dynamics and Molecular Motion. Yet, even those can't solve protein folding reliably, and take months on a super computer. Pattern recognition, that is seeing similar motifs and the like on the fly, is something only the human brain can really do, and is why it is superior to computers when solving protein folds. It's just giving us a graphical, manipulable interface, with computational feedback, suddenly makes solving these problems accessible; and that's what Foldit has provided.


Odd
By Matthiasa on 9/19/2011 8:39:18 PM , Rating: 4
2008 + 3 weeks = 2011... something seems off there.




RE: Odd
By geddarkstorm on 9/20/2011 12:11:06 PM , Rating: 2
Foldit's been around for awhile, they just haven't thrown something new at the gamers like this till now. That, and it takes time to verify the result, using it as guidance in the verification experiments.


Ender's Game!
By mfenn on 9/19/2011 5:00:32 PM , Rating: 3
Take that you buggers!




RE: Ender's Game!
By Camikazi on 9/19/2011 9:01:43 PM , Rating: 2
Ender, I should reread that series, I loved it :D


How to verify result(s) ?
By Gondor on 9/20/2011 3:35:05 AM , Rating: 2
So they created this game and gamers came up with correct solution. Who was it to confirm that John Doe actually got the right one ?

Was it computer, which according to the article can't do the folding because it 'lacks spational reasoning skills', or did these gamers submit three years worth of results of their gameplay (millions and more, no doubt) to some researcher(s) to verify ?




RE: How to verify result(s) ?
By zephyrprime on 9/20/2011 11:32:34 AM , Rating: 2
The computer probably verified it immediately after the folding plan was first generated by the gamer. The program does have spatial physics comprehension of the task at hand (protein folding).


"It seems as though my state-funded math degree has failed me. Let the lashings commence." -- DailyTech Editor-in-Chief Kristopher Kubicki














botimage
Copyright 2014 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki