[jdp]

http://video.google.com/videoplay?docid ... 0976635143

[Thomas An.]

Very clever

[x_site]

wow... beats playing solitaire....

[tom]

Prodigious!

Btw see in the comments, one of my clone has asked a good question... LOL

[tokiop]

wow madly brilliant... the human knowledge and cognition transfert on silicium is incredibly growing...

imagine global 3d modelisation with peakaboo mask technique..applications of this concept are infinite...

[Mihai]

I don't know.....somehow I don't feel so amazed by this. I mean with the web constantly changing, half the images on google are gone when you click them.

The common sense game seemed more interesting for AI development.

[tokiop]

sure, it is just the begining, the last games are naturally more developped!

[jdp]

I don't know.....somehow I don't feel so amazed by this. I mean with the web constantly changing, half the images on google are gone when you click them.

The common sense game seemed more interesting for AI development.

well I guess that the final goal is to improve the AI algorithm to recognize things represented in the image, not to classify the images itself.
What this guy did is to leave a part the math for a moment going back to human science, thinking about the interaction within different kind of intelligence instead of the math involved in it.

To be honest this is what I found really interesting: using the interaction within different fields, classic science and human science, to analyze and explain complexity.

[b-kandor]

In a similiar vein, I read recently about the The Global Consciousness Project :

Apparently this shows a correlation between random event generator (running on computers) and significant world events traumatic or overwise. It's interesting.

"To study the effects of a possible global consciousness, we have created a world-spanning network of devices sensitive to coherence and resonance in the mental domain. Continuous streams of data are sent over the internet to be archived and correlated with events that may evoke a world-wide consciousness. Examples that appear to have done so include both peaceful gatherings and disasters: a few minutes around midnight on any New Years Eve, the first hour of NATO bombing in Yugoslavia, the Papal visit to Israel, a variety of global meditations, several major earthquakes, and September 11 2001. "

[daros]

I don't think this guy is a genius and i think that his indexing technique will give mediocre search results.
Until they want to index dog images may it's ok. But try to index the dog's race or the tree's scientific name the system will show immediatly his lacks because it's obvius that only a 0.1% of peolple knows the scientific name of a tree.
I Think that this indexing technique will handle true knowledge as an "exeption" and expetions are always removed in a statistical analisys.
The short time allowed to input the data excludes any research operation with the warranty af mediocre input.
In my opinion 99.9% of WWW content is useless, a monument to common sense, and i'ts even more useless to put the whole umanity to confirm the common sense of that content.
Sorry for this for this negative comment, but todays is sunday and i'm still working in the office

[Thomas An.]

The fascinating thing for me is the *concept* of energy recycling (human thought being a form of energy in this sense).

... in much the same way that, from an engineering standpoint, the automatic winding mechanisms of early wristwatches (the ones with a built-in counter weight that captures the motions of your hand) is also *immensely fascinating* (the idea of energy redistribution)... but not many of my friends see what is the big deal with that. (I suppose it is a geeky thing to look at it that way).

Until they want to index dog images may it's ok. But try to index the dog's race or the tree's scientific name the system will show immediatly his lacks because it's obvius that only a 0.1% of peolple knows the scientific name of a tree.

Isn't that what he was talking about, when mentioning the "taboo" words ? after a period of time some words of the picture are considered taboo words by the game and you cannot use them any more ... this way the game becomes more difficult and requires more specialized knowledge. So, you can no longer say "dog" ... you are forced to say "Doberman" ... or Bristlecone pine, instead of tree ...

[DELETED]

DELETED

[DELETED]

DELETED

[daros]

Man
black and white
camera
film
movie
glove
leather
cap
nose
hand
outdoor
ear
cannon jack
open mounth
old photo
telescope
35mm
Arriflex Panavision or watever
camera man
Abel Ferrara
Director
the camera model

I want only say that before this image is indexed in an interesting way you need that ESP chose it may 100 times in a ocean of 1.000.000.000 images and assign it casually to two people in a ocean of 100.000.000 that indentify Abel Ferrara.
Since you are not allowed to do research the label "abel ferrara" will stay blank until the next ice age.

May i completly wrong thomas but this is the feeling i have.

[daros]

in this well known example the problem points to his own paradigm


only the content creator is able to give the correct label.

i'm joking... by sure it's possible to reach something interesting with collaborative knowledge + databases but the ESP game has for me a bit disappointing approach.