Google Blogoscoped

Forum

Mutating Pictures Video & Faces of Facemaker  (View post)

Hong Xiaowan [PersonRank 10]

Friday, December 12, 2008
11 years ago2,522 views

So funny

Veky [PersonRank 10]

11 years ago #

> It’s interesting to note that the speed of the evolution in this experiment is not dependent on time, but rather on the number of people participating. In the first couple of days when many people participated, picture pool generations passed by quickly leading to many mutations and a fast face evolution (often the oldest living member in the pool of 1000 pictures was mere minutes old). After the first couple of days, traffic nearly died down, and so evolution came to a halt. The first days of the experiment created more power than the whole following year taken together <

Are you sure? To me, a more plausible explanation would be that blind mutations can really significantly alter the image for the better easily only in the first stages, when the images are much more different than faces. Later, when images start to resemble faces, it's very hard to find a suitable mutation to elicit a positive vote from people participating, regardless of how many of them are there.

I suggest you repeat the experiment, this time with already evolved images, and try to attract as much visitors as you can. My hypothesis is that you won't notice a significant improvement in resemblance to faces.

Philipp Lenssen [PersonRank 10]

11 years ago #

> Later, when images start to resemble faces, it's very hard
> to find a suitable mutation to elicit a positive vote from
> people participating, regardless of how many of them are there.

I was more referring to the new pools I created than to the face pool. I don't have proof either way, all I know is that during the peak time there were e.g. over 100,000 views on the page whereas some weeks later the views wore more like dozens a day – and I know that with dozens a day no evolution at all can happen (well "none at all" in pragmatic terms... of course there's a tiny fraction of evolution happening with every single rating). The new pools I created after some days – like the body and animal pool – thus didn't receive enough power to test if they would have mutated well. I think these pools were harder (lack of symmetry for instance, or our ability to see faces in almost everything!) but I think they too might've evolved into something interesting with more power.

As for the face pool, you're right, that one might well nearly stagnate at a certain point using the 0-10 rating system (and perhaps at a later point, also using the image comparison approach of Facemaker). Or the votes vs evolution-speed factor might have decreased to an unrealistic point where it would take more and more votes to cause even slight changes.

(There may be another side to this: that, if a pool is not evolving quickly enough, it cannot help keep people's interest, which in turn makes people leave and then let the pool completely die off...)

This thread is locked as it's old... but you can create a new thread in the forum. 

Forum home

Advertisement

 
Blog  |  Forum     more >> Archive | Feed | Google's blogs | About
Advertisement

 

This site unofficially covers Google™ and more with some rights reserved. Join our forum!