Embedded Files
Mathematics of Interestingness
Mathematics of Interestingness
In this presentation, I propose a mathematical framework wherein "interestingness" is defined quantitatively as the absolute derivative of predictability over time (i.e., changes in predictability are interesting). Furthermore, "curiosity" is defined as the expected value of future interestingness. It is shown how the framework explains the interestingness of broken symmetry in visual patterns, including the famous Mona Lisa "smile."
I then discuss how we experience interestingness over a range of human expressions, including music, storytelling, episodic television, athleticism, intoxication, sensory immersion, socio-political, and conceptual art.
Algorithmic Creativity
Algorithmic Creativity
Next, I propose a generalized algorithm for producing interesting works of art, and discuss the effects of the underlying computing strategy and/or optimization criteria on its results. I examine the concept with reference to the works of M.C. Escher, the Spice Girls, naked mole rats, and the Great Pyramid of Giza.
The presentation concludes with recommendations for human and AI-driven creators to produce more interesting art. This work was originally presented April 3rd, 2010 at the Gray Area Foundation for the Arts in San Francisco.
Interestingness-GAFFTA-2010.pdfPage updated
Google Sites
Report abuse