Saturday, 23 March 2013

Jürgen Schmidhuber: Low-complexity art and more!

Jürgen Schmidhuber is arguably one of the world’s most interesting researchers in AI. He is a computer scientist and artist known for his work on machine learning, Artificial Intelligence (AI), artificial neural networks, digital physics, and low-complexity art. Schmidhuber is co-director of the Swiss AI lab IDSIA in Lugano and a professor of Cognitive Robotics at the Tech University Munich. It is reported that since he was 15 years old, his main scientific ambition has been to build an optimal scientist, then retire! This is the driving force behind his research on self-improving Artificial Intelligence. Between 2009-2012, the recurrent neural networks and deep feed-forward neural networks developed in his research group have won eight international competitions in pattern recognition and machine learning. His formal theory of creativity & curiosity & fun (1990-2010) explains art, science, music, and humor. Yes, his CV is impressive, but what caught my attention in particular was his 'low-complexity art' which is reinforced and mentioned in his formal theory of creativity & curiosity & fun. 

In relation to Low-complexity art: Schmidhuber explains that when representing an object, many artists attempt to convey its 'essence'. In an attempt to formalize certain aspects of depicting the essence of objects, Schmidhuber puts forward an art form called low-complexity art which can be seen as the 'computer age equivalent' to minimal art. "Its goals are based on concepts from algorithmic information theory. A low-complexity artwork can be specified by a computer algorithm and should comply with two properties: (1) the drawing should ``look right," and (2) the Kolmogorov complexity of the drawing should be small (the algorithm should be short) and a typical observer should be able to see this. Examples of low-complexity art are given in the form of algorithmically simple cartoons of various objects. Attempts are made to relate the formalism of the theory of minimum description length to informal notions such as ``good artistic style" and 'beauty'". 

Low complexity art is explained perfectly in relation to the image above as Schmidhuber explains that:

"Many observers report they derive pleasure from discovering simple but novel patterns while actively scanning this self-similar Femme Fractale (image above). The observer's learning process causes a reduction of the subjective complexity of the data, yielding a temporarily high derivative of subjective beauty: a temporarily steep learning curve. Similarly, the computer-aided artist got reward for discovering a satisfactory way of using fractal circles to create this low-complexity artwork, although it took him a long time and thousands of frustrating trials. Here is the explanation of the artwork's low algorithmic complexity:- The frame is a circle; its leftmost point is the center of another circle of the same size. Wherever two circles of equal size touch or intersect are centers of two more circles with equal and half size, respectively. Each line of the drawing is a segment of some circle, its endpoints are where circles touch or intersect. There are few big circles and many small ones. This can be used to encode the image very efficiently through a very short program. That is, the Femme Fractale has very low algorithmic information or Kolmogorov complexity."

Below is a video of Schmidhuber giving a talk on his Algorithmic Theory of creativity & curiosity & fun  that explains art, science, music, and humor at the Singularity summit 2009 in NY. 

Schmidhuber is a truly fascinating character. You can read more about low-complexity art here and Schmidhuber's other work through the links provided above.