In response to Shh's complaint (some time ago now) that there were never any aesthetics threads in here, I thought I'd ask everyone what they make of Schmidhuber's algorithmic theory of beauty.

So, here's the postulate:

Among several patterns classified as "comparable" by some subjective observer, the subjectively most beautiful is the one with the simplest (shortest) description, given the observer's particular method for encoding and memorizing it.

Taken from: http://www.idsia.ch/~juergen/beauty.html Which also contains the link to this PDF summary of Schmidhuber's ideas: http://www.idsia.ch/~juergen/sice2009.pdf

In this summary of previous work, I argue that data becomes temporarily interesting by itself to some self improving, but computationally limited, subjective observer once he learns to predict or compress the data in a better way, thus making it subjectively more “beautiful.” Curiosity is the desire to create or discover more non-random, non-arbitrary, “truly novel,” regular data that allows for compression progress because its regularity was not yet known. This drive maximizes “interestingness,” the first derivative of subjective beauty or compressibility, that is, the steepness of the learning curve. It motivates exploring infants, pure mathematicians, composers, artists, dancers, comedians, yourself, and recent artificial systems.

Here's some excerpts from a lecture by Schmidthuber on his ideas on this and other things from the Singularity Summit last year:

[youtube]http://www.youtube.com/watch?v=Ipomu0MLFaI[/youtube]

And a wiki entry:

Jürgen Schmidhuber described an algorithmic theory of beauty which takes the subjectivity of the observer into account and postulates: among several observations classified as comparable by a given subjective observer, the aesthetically most pleasing one is the one with the shortest description, given the observer’s previous knowledge and his particular method for encoding the data[39][40]. This is closely related to the principles of algorithmic information theory and minimum description length. One of his examples: mathematicians enjoy simple proofs with a short description in their formal language. Another very concrete example describes an aesthetically pleasing human face whose proportions can be described by very few bits of information[41][42], drawing inspiration from less detailed 15th century proportion studies by Leonardo da Vinci and Albrecht Dürer. Schmidhuber's theory explicitly distinguishes between what's beautiful and what's interesting, stating that interestingness corresponds to the first derivative of subjectively perceived beauty. Here the premise is that any observer continually tries to improve the predictability and compressibility of the observations by discovering regularities such as repetitions and symmetries and fractal self-similarity. Whenever the observer's learning process (which may be a predictive neural network - see also Neuroesthetics) leads to improved data compression such that the observation sequence can be described by fewer bits than before, the temporary interestingness of the data corresponds to the number of saved bits. This compression progress is proportional to the observer's internal reward , also called curiosity reward. A reinforcement learning algorithm is used to maximize future expected reward by learning to execute action sequences that cause additional interesting input data with yet unknown but learnable predictability or regularity. The principles can be implemented on artificial agents which then exhibit a form of artificial curiosity.

http://en.wikipedia.org/wiki/Aesthetics ... aesthetics

Let's see where this goes.

"Don't forget to tip your waitresses! Good night."

He's funny. Interesting topic. I remember reading about a study done on beauty, in Australia, I believe where human faces were judged by a rather large test group. The conclusion was that average was most beautiful. Seems that would fit with the idea of simplest to encode. Interesting topic

chippy wrote:"Don't forget to tip your waitresses! Good night."

He's funny. Interesting topic. I remember reading about a study done on beauty, in Australia, I believe where human faces were judged by a rather large test group. The conclusion was that average was most beautiful. Seems that would fit with the idea of simplest to encode. Interesting topic

I recommend the PDF above to anyone with a lay interest in the subject, as the author has elected to leave the formal explanations within their original papers, making for very clear and enjoyable reading on a variety of subjects related to his theory.

edited to add that the full version of Schmidhuber's lecture is available in several parts, beginning here:

[youtube]http://www.youtube.com/watch?v=sa65t1fVsNA[/youtube]

Schmidhuber claims to have answered, via a simple algorithm, aesthetic questions of 'what is beauty', 'what is interesting' and 'what is curiosity' - with implications that extend into ethics and the philosphy of consciousness, such that they would be relevant to at least three of the topics in the top ten of this forum, and no-one, not even the hardened 'nonnies' of the forum, have anything to say about it, given more than twenty-four hours to read and digest what he is saying?

I have to say I am surprised, to say the least!

Jef wrote:Schmidhuber claims to have answered, via a simple algorithm, aesthetic questions of 'what is beauty'

Well, this should allow him to automatically create works of art of great beauty. Where can I see or hear those works?

'what is interesting' and 'what is curiosity' - with implications that extend into ethics and the philosphy of consciousness, such that they would be relevant to at least three of the topics in the top ten of this forum, and no-one, not even the hardened 'nonnies' of the forum, have anything to say about it, given more than twenty-four hours to read and digest what he is saying?

I have to say I am surprised, to say the least!

michael^3 wrote:

Jef wrote:Schmidhuber claims to have answered, via a simple algorithm, aesthetic questions of 'what is beauty'

Well, this should allow him to automatically create works of art of great beauty.

No, this ignores the strong subjective element within the theory. Because, what is subjectively more beautiful to each observer depends upon their particular method for encoding and memorizing, which, in turn, depends upon what they have previously memorized and encoded. So, perhaps the strongest claim, in this regard, that can be made from the theory is that, if provided with sufficient information about a subjects previous experience, it can be predicted, from a class of things identified as comparable by the subject, which will be considered more intrinsically beautiful. It follows that if sufficient information is provided about the subject, they can be provided with a 'tailor-made' article of beauty - but mileage varies, and the notion that there is some objective concept of the beautiful is rejected.

To emphasise this point as I see it, it is one of the strengths of this theory that it adopts the subjective history of the observer as a variable, and rejects the idea of beauty as objective.

Secondly, Schmidhuber is quite clear on the point that his theory deals only with intrinsic, or 'pure', beauty, and does not take account of external influences upon the subjective history, such as memories of a things as being comforting, or providing a sense of safety. The theory pertains purely to the aesthetics, in an 'art for art's sake' manner, to discover what you might call the intrinsic value of beauty. By ignoring the extrinsic values placed upon art, the 'how you feel about it' sensation about the things presented to a subject can vary, despite that the intrinsic beauty is precisely accounted for.

Finally, I would like to draw your attention to the symmetries drawn out by Schmidthuber between the artist and the observer of the art, and the similarity in the compression achieved by both, as an explanation of why it is that no artist is universally capable of satisfying an entire audience. Artists communicate that which it is compressively satisfying for them to communicate to those to whom it is compressively satisfying to receive such communication. In creating art, the artist conceives of an audience which will recieve a similar satisfaction in appreciation of the art as the artist did in creating it. So, any art created by Schidthuber would be satisfying to the observer in degrees which can be measured in reference to the satisfaction given to someone with the subjective history of Schmitgruber. Artists create for their own satisfaction, (notwithstanding the externality of public approval, which is considered), and are appreciated as beautiful by those who share this, particular, satisfaction. There is no universailty of beauty beyond the shared experience.

My apologies for any long-windedness or incoherence in the above... I'm really quite drunk, and shall endeavour to either correct or confirm what I write this evening, tomorrow.

Where can I see or hear those works?

The argument would run that those works you already consider to be beautiful would fit within the parameters of the theory.