The new forums will be named Coin Return (based on the most recent vote)! You can check on the status and timeline of the transition to the new forums here.
The Guiding Principles and New Rules document is now in effect.
LttP: Human Computation and Games With a Purpose
CaspianX
Registered User regular
Registered User regular
This is nearly two years old, and yet I'm just now finding out about it, and I'm so amazed and impressed I had to share it here, just in case there are some people here who, like me, missed out on it the first time around.
The story begins with a Carnegie Mellon University PHD by the name of Luis Von Ahn, and a Google TechTalks presentation he gave in July 2006. The 51-minute presentation is here for those interested in watching it:
http://www.youtube.com/watch?v=qlzM3zcd-lk
I'd recommend you give it a look, especially if you're interested in game design, software design, pattern recognition, artificial intelligence, or especially clever alternate uses for gamer input. However, even if none of that stuff really drives you, you still may wanna' check it out - it's all delivered in common terms that even less tech-inclined folks should be able to follow.
However, the gist of it is this - there are some things that computers can't do very well that the human brain can do very easily. Describing the contents of a picture, finding an object in an image just from a description, stuff like that. These are things that have a very practical use - for example, image searches would be a lot more accurate if images were tagged with more descriptive tags. Up until Von Ahn came along, companies like Google and Flikr actually hired people to write these descriptors. However, he came up with a clever way to get people to want to do it willingly.
The Human Computation concept is to use people as an advanced computation network of sorts, making use of collective brainpower to do things that computers can't. As written about in this post on Sirlin.net, this is something that has already proven to be an extremely potent force to be reckoned with, solving incredibly complex problems in ARGs at amazing speeds, and it only requires a clever device to turn this massive force towards a productive goal.
Enter Games With a Purpose, programs that are surprisingly fun and addictive (this ain't Edutainment we're talking about here), and serve a practical purpose. Here's what I'm talking about:
The ESP Game
The ESP Game is a Java-based web game where two anonymous players each look at the same image and try to come up with the same word to describe it, and as quickly as possible. An image featuring a dog playing in the park will attract descriptors like dog, park, grass, sky, et cetera, and players generally try to put in many words as fast as they can in an attempt to match their partner. Words commonly used by previous players are added to a "taboo" list, making them off-limits (requiring a bit more ingenuity with subsequent play-throughs). It's fun, it's easy, it's addictive, and in the background, it's making those labels for search engines.
Peekaboom
Peekaboom takes the basic concept of the ESP Game a step further. Here we have images and we have labels, but what if we wanted to get the computer to recognize what part of the image these words describe? Peekaboom addresses this by having players take turns as "Peek" and "Boom". Boom is given an image and a word, and they must get Peek to guess what that word is. They do this by revealing parts of that image to Peek, and giving other hints to try and get Peek to guess it correctly. Again, fun and addictive. Meanwhile, the parts of the image they choose to reveal tell the computer "this is the part of the image that's the dog".
Verbosity and Phetch
I haven't had as much time to look into these two, but they're from the same people, so I figgered I'd link 'em here just in case anyone's interested.
Anyways, I'll be curious to hear your thoughts on this stuff.
Edit: Damn. Can't get Verbosity to work. It keeps saying it can't connect to the server.
The story begins with a Carnegie Mellon University PHD by the name of Luis Von Ahn, and a Google TechTalks presentation he gave in July 2006. The 51-minute presentation is here for those interested in watching it:
http://www.youtube.com/watch?v=qlzM3zcd-lk
I'd recommend you give it a look, especially if you're interested in game design, software design, pattern recognition, artificial intelligence, or especially clever alternate uses for gamer input. However, even if none of that stuff really drives you, you still may wanna' check it out - it's all delivered in common terms that even less tech-inclined folks should be able to follow.
However, the gist of it is this - there are some things that computers can't do very well that the human brain can do very easily. Describing the contents of a picture, finding an object in an image just from a description, stuff like that. These are things that have a very practical use - for example, image searches would be a lot more accurate if images were tagged with more descriptive tags. Up until Von Ahn came along, companies like Google and Flikr actually hired people to write these descriptors. However, he came up with a clever way to get people to want to do it willingly.
The Human Computation concept is to use people as an advanced computation network of sorts, making use of collective brainpower to do things that computers can't. As written about in this post on Sirlin.net, this is something that has already proven to be an extremely potent force to be reckoned with, solving incredibly complex problems in ARGs at amazing speeds, and it only requires a clever device to turn this massive force towards a productive goal.
Enter Games With a Purpose, programs that are surprisingly fun and addictive (this ain't Edutainment we're talking about here), and serve a practical purpose. Here's what I'm talking about:
The ESP Game
The ESP Game is a Java-based web game where two anonymous players each look at the same image and try to come up with the same word to describe it, and as quickly as possible. An image featuring a dog playing in the park will attract descriptors like dog, park, grass, sky, et cetera, and players generally try to put in many words as fast as they can in an attempt to match their partner. Words commonly used by previous players are added to a "taboo" list, making them off-limits (requiring a bit more ingenuity with subsequent play-throughs). It's fun, it's easy, it's addictive, and in the background, it's making those labels for search engines.
Peekaboom
Peekaboom takes the basic concept of the ESP Game a step further. Here we have images and we have labels, but what if we wanted to get the computer to recognize what part of the image these words describe? Peekaboom addresses this by having players take turns as "Peek" and "Boom". Boom is given an image and a word, and they must get Peek to guess what that word is. They do this by revealing parts of that image to Peek, and giving other hints to try and get Peek to guess it correctly. Again, fun and addictive. Meanwhile, the parts of the image they choose to reveal tell the computer "this is the part of the image that's the dog".
Verbosity and Phetch
I haven't had as much time to look into these two, but they're from the same people, so I figgered I'd link 'em here just in case anyone's interested.
Anyways, I'll be curious to hear your thoughts on this stuff.
Edit: Damn. Can't get Verbosity to work. It keeps saying it can't connect to the server.
CaspianX on
0
Posts
Yeah, despite that the guy has a bit of an odd accent, he's a pretty good presenter.