Ruben Van Havermaet's Library tagged AI → View Popular

Filter:
All
Annotated

30 Jul 09

Penousal Machado's publication

eden.dei.uc.pt/~machado - Preview

art artscience newmedia AI music

29 Jul 09

28 Apr 09

Harold Cohen

features a list of key articles by Harold Cohen

crca.ucsd.edu/~hcohen - Preview

art AI newmedia painting

22 Mar 08

Why Lisp?

nostoc.stanford.edu/...whylisp.html - Preview

ai computer science lisp

true macros, that are
critical when you are inventing a new programming language
most natural language in which to write complex 'symbolic
reasoning'
1 more annotations...
- involves manipulating conceptual
  objects

Self-organizing map - Wikipedia, the free encyclopedia

en.wikipedia.org/...Kohonen_maps - Preview

ai computer science learning

unsupervised learning
useful for visualizing low-dimensional views of high-dimensional data
21 more annotations...
- operate in two modes: training and mapping
- consists of components called nodes or neurons
- Associated with each node is a weight vector of the same dimension as the input data vectors and a position in the map space
- describes a mapping from a higher dimensional input space to a lower dimensional map space
- small number of nodes behave in a way that is similar to K-means
- U-matrix value of a particular node is the average distance between the node and its closest neighbors
- Large SOM display properties which are emergent
- goal of learning in the self-organizing map is to cause different parts of the network to respond similarly to certain input patterns
- weights
- sampled evenly from the subspace spanned by the two largest principal component eigenvectors
- much faster because the initial weights already give good approximation of SOM weights
- training example is fed to the network, its
- Euclidean distance to all weight vectors is computed
- weights of the BMU and neurons close to it in the SOM lattice are adjusted towards the input vector
- best matching unit
- Wv(t + 1) = Wv(t) + Θ (v, t) α(t)(D(t) - Wv(t)),
- monotonically decreasing learning coefficient
- At the beginning when the neighborhood is broad, the self-organizing takes place on the global scale. When the neighborhood has shrunk to just a couple of neurons the weights are converging to local estimates
- any kind of object which can be represented digitally and which has an appropriate distance measure associated with it and in which the necessary operations for training are possible can be used to construct a self-organizing map
- Randomize the map's nodes' weight vectors
  Grab an input vector
  Traverse each node in the map
  
  Use Euclidean distance formula to find similarity between the input vector and the map's node's weight vector
  Track the node that produces the smallest distance (this node is the best matching unit, BMU)
  Update the nodes in the neighbourhood of BMU by pulling them closer to the input vector
  
  Wv(t + 1) = Wv(t) + Θ(t)α(t)(D(t) - Wv(t))
  Increment t and repeat while t < λ
- SOM forms a semantic map where similar samples are mapped close together and dissimilar apart

08 Feb 08

Neural Networks - A Systematic Introduction

a free neural network book available online

page.mi.fu-berlin.de/...index.html - Preview

AI computerscience ebooks learning neuralnetworks research

02 Dec 07

Autopoesis

christianhubert.com/...Autopoesis.html - Preview

AI artscience autopoesis newmedia

22 Nov 07

Visualization and art

www.cscs.ch/...VizArt.html - Preview

AI art artscience newmedia

see the unseen and communicate it
- creativity
- intuition
- willingness to throw out the rules
- close observation of nature
- ability to see patterns where others see chaos
43 more annotations...
- common interests and themes:
- attern recognition
- unbiased eye
- details) are not important and should be ignored
- Visual literacy
- learn to see
- Artists benefit from
  rapid test of new ideas and from multiple approaches to the same problem
- A work of art always remains open for interpretation
- Scientists
- communicate an interpretation that embodies testable content in an
  unambiguous way
- Arts integrate, specialized science disintegrate knowledge
- ultimate goals of scientific models
- objectivity and reproducibility
- rarely allude to this psychical process
- deny the motivation and inspiration behind this creative visual experience
- Visualization
- information hidden inside the data
- leverages human perception to make structure, regularities and patterns apparent
- mental models to external cognitive artifacts
- anchors
- communication tool
- exploratory tool
- If we trace out what we behold and experience through the language of logic we are doing science;
  if we show it in forms whose interrelationships are not accessible to our conscious thought but
  are intuitively recognized as meaningful, we are doing art
- fidelity
- part of a process of discovery
- aid the
  intuition in identifying relationships
- representing
  all data using the same numerical code, computers make it easy to map one representation into another
- Most mappings in both science and art go from non-visual media to visual media.
  Is it possible to create mappings that will go into the opposite direction
- modernism reduced the
  particular to its Platonic schemas
- data
  visualization as a new abstraction
- quantitative data is reduced to its patterns and structures that are then exploded into many rich and concrete
  visual images
- This promise makes data mapping into the
  exact opposite of the Romantic art concerned with the sublime. In contrast, data visualization art is concerned
  with the anti-sublime. If Romantic artists thought of certain phenomena and effects as un-representable, as
  something which goes beyond the limits of human senses and reason, data visualization artists aim at precisely
  the opposite: to map such phenomena into a representation whose scale is comparable to the scales of human
  perception and cognition.
- Analogy and metaphor
- exploration of ideas,
  knowledge, and options
- Breaking with convention
- total immersion in the activity
- embraced by taking a holistic, or systemic,
  view
- different
  approaches and viewpoints active at the same time
- Parallel channels
- Modeling the creative process support systems shows that knowledge and visualization
  are essential ingredients of creativity work
- The disappointing fact is that the majority of design is not in groundbreaking innovation,
  but in variations on something successful
  - cf. Avant-Garde as Software: in the post-industrial information age instead of focusing on media production, there's an emphasis on recombining existing media (pastiche, collage, ...)
    cf. Cultural Change (F.Jameson) - on 2007-11-22
  Add Sticky Note
- generating pre-inventive forms or structures and then tweaking (that's jargon for twisting,
  forcing, inquiring, etc.) the form to relate it to a novel avenue of thought or procedure
- Artist: how much objectivity before starts losing intuition?
  Scientist: how much intuition before they lose scientific foundations?