XPost: comp.ai.neural-nets, comp.graphics.algorithms   
   From: "a_jakulin@"@hotmail.com   
      
   Gernot Hoffmann wrote:   
   > I think your approach has really nothing to do with aesthetics.   
   > But with artificial intelligence - find plausible interpretations   
   > input data, based on some assumptions like 'isolated line segments   
   > should be connected' or 'nearly parallel lines should be parallel',   
   > or 'nearly straight lines should be straight lines'.   
      
   You are quite right. Being situated in AI, and putting 15% of the work   
   into aesthetic considerations makes me feel that it is mostly   
   aesthetics, but when you go out of your little field, you're quickly   
   reminded otherwise. That's the bane of interdisciplinarity.   
      
   Aesthetics is a much wider term than what we were doing. Indeed, we   
   were doing pattern recognition. Finding plausible explanations for   
   data can be seen as pattern recognition. With the Bayes rule:   
      
   P(Explanation | Data) = P(Explanation) P(Data | Explanation) / P(Data)   
      
   P(Explanation) - prior   
   P(Explanation | Data) - posterior   
   P(Data | Explanation) - likelihood of the explanation   
   P(Data) - evidence   
      
   The prior - P(Explanation) - is the key here. It must be constructed,   
   and it may be subjective. It is the formalization of the assumptions   
   you list above. Therefore, pattern recognition cannot be independent   
   from an aesthetically-informed prior when the Explanations are   
   diagrams.   
      
   So, pattern recognition can be aesthetically-informed. Of course this   
   is narrower than aesthetics as a whole.   
      
   > 'Informations-Aesthetik' was a major issue about 40 years ago, when   
   > computer graphics started.   
   > Key words for a Google Search: 'Max Bense', 'Rul Gunzenhaeuser'.   
      
   Very useful - thanks. Max Bense defines aesthetics through information   
   theory:   
   aesthetics = redundancy / information   
      
   Anyway, my previous knowledge about aesthetics was based on how it's   
   presented in information visualization, not in computer graphics. For   
   example, see "Graph Drawing:past - present - future"   
   (http://www.csse.monash.edu.au/~gfarr/research/GraphDrawing02-Mel.ppt)   
      
   --   
   mag. Aleks Jakulin   
   http://www.ailab.si/aleks/   
   Artificial Intelligence Laboratory,   
   Faculty of Computer and Information Science,   
   University of Ljubljana, Slovenia.   
      
   [ comp.ai is moderated.  To submit, just post and be patient, or if ]   
   [ that fails mail your article to , and ]   
   [ ask your news administrator to fix the problems with your system. ]   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)