Neural aesthetics

From Monoskop
Revision as of 23:07, 28 November 2018 by Dusan (talk | contribs)
Jump to navigation Jump to search
Partial results of the "hello world" test of a machine learning algorithm, automated recognition of handwritten digits, showing 125 test cases that the network got wrong. Each case is labeled by the network’s guess. The true classes are arranged in standard scan order. Source: Hinton et al 2006.
Basic structure of a neural network. Several techno-logical forms can be identified in the concept: scansion, that is discretisation or digitisation since the age of radio, TV, etc.; feedback loop, or the basic concept of cybernetics; and network form, here inspired by biological neurons. Diagram by Matteo Pasquinelli with Lukas Rehm, 2017. Source.

A resource on recent work between art/design and artificial neural networks in machine learning.

Related notions: AI art, creative AI, art and artificial intelligence.

Events

2015

  • The Lab at the Google Cultural Institute, Paris, launches a 'machine learning for art' residency, early 2015 - mid-2016. Artists: Mario Klingemann, Cyril Diagne. Talk. Works.
  • DeepDream launched by Google's software engineers Alexander Mordvintsev, Christopher Olah and Mike Tyka, 17 Jun 2015. Reddit post from a day earlier. Vice coverage. Source code.
  • (Artifical Intelligence) Digitale Demenz exhibition, HMKV, Dortmund, 14 Nov 2015-6 Mar 2016. An exhibition exploring the relationship between contemporary art and artificial intelligence. Works by Erik Bünger, John Cale, Brendan Howell, Chris Marker, Julien Prévieux, Suzanne Treister, and !Mediengruppe Bitnik. Curated by Thibaut de Ruyter.

2016

2017

2018

Artists, designers

Initiatives

Literature, data, resources

Art practice and criticism

Online galleries and collections

See also exhibitions in the 'Events' section above.

Recent debates on artificial intelligence in the humanities and social sciences

Software

  • Magenta Studio, a suite of free music-making tools using Magenta's machine learning models. Available as an Ableton plugin or as standalone Electron apps. Launched Nov 2018.
  • Magenta demos

Data sets

Courses and textbooks for artists

Textbooks
Video lectures
  • Gene Kogan, "Machine Learning for Artists", course taught at ITP, New York University, New York, Spring 2016. 6 lectures, ~15 hours.
  • Gene Kogan, "The Neural Aesthetic", course taught at School Of Machines, Making, and Make-Believe, Berlin, Summer 2016. 8 lectures, ~18 hours.
Introductions
MOOC
Resources
  • A.I. Experiments, a resource on machine learning art developed by Alexander Chen, creative director at Google Creative Lab, with Yotam Mann, C Christiansen, Jonas Jongejan, Gene Kogan, Kyle McDonald, et al. Launched on 16 Nov 2016.
  • People + AI Research, Google Design, 2018ff.

Research papers

Scientific introduction into deep learning

Historization of deep learning