Difference between revisions of "Self-Organizing Systems Literature Recommendations"
(→Introduction) |
(→List) |
||
| Line 11: | Line 11: | ||
# [http://gral.ip.rm.cnr.it/Pubblicazioni/English/Book%20Chapters/marocco_sab9.pdf Origins of communication in evolving robots]. Marocco, Nolfi - Lecture Notes in Computer Science, 2006 - Springer. This is about grounding communication. Robots evolve neural network weights such that the speaker-microphone interactions between robots becomes meaningful. That is they tell each other if they are on a colored spot. It is self-organized for the reason that meaningful communication elements are not predefined but arise by interactions between the robots. (They also use the term in the title of some of their articles. See also Steels etc.) | # [http://gral.ip.rm.cnr.it/Pubblicazioni/English/Book%20Chapters/marocco_sab9.pdf Origins of communication in evolving robots]. Marocco, Nolfi - Lecture Notes in Computer Science, 2006 - Springer. This is about grounding communication. Robots evolve neural network weights such that the speaker-microphone interactions between robots becomes meaningful. That is they tell each other if they are on a colored spot. It is self-organized for the reason that meaningful communication elements are not predefined but arise by interactions between the robots. (They also use the term in the title of some of their articles. See also Steels etc.) | ||
| + | # [http://www.ifi.uzh.ch/ailab/teaching/NN2007/art/art_paper.pdf The ART of adaptive pattern recognition by a self-organizing neural network] Carpenter, Grossberg - Computer, 1988. This article describes a class of neural networks in which there is self-organization in how patterns are stored. A neural network is plastic, so after enough new input patterns, the old information is washed away. Adaptive resonance theory [ART] enables a network to be robust against arbitrary sequences of inputs in a self-regulated manner. It is self-organized in the sense that it is unsupervised. | ||
Revision as of 11:08, 17 July 2009
Introduction
This section contains recommendations of literature by workshop participants. It is not necessarily work done by the participants, but rather the type of literature that is recommended to the other to become more versed in the topic of self-organizing systems. Books that are recommended might be likely used in the case a master of self-organizing systems will be set up. If possible, a text explains why a book or article is about self-organizing systems or interesting as background material. For example graph theory as such does not necessarily describe self-organization. An explanation that relates those will be clarifying.
(Remove this later: For now, say every person can add 3 references to books (link to Amazon) and/or articles (link to web) that are interesting).
List
Marc Timme recommends:
Anne van Rossum recommends:
- Origins of communication in evolving robots. Marocco, Nolfi - Lecture Notes in Computer Science, 2006 - Springer. This is about grounding communication. Robots evolve neural network weights such that the speaker-microphone interactions between robots becomes meaningful. That is they tell each other if they are on a colored spot. It is self-organized for the reason that meaningful communication elements are not predefined but arise by interactions between the robots. (They also use the term in the title of some of their articles. See also Steels etc.)
- The ART of adaptive pattern recognition by a self-organizing neural network Carpenter, Grossberg - Computer, 1988. This article describes a class of neural networks in which there is self-organization in how patterns are stored. A neural network is plastic, so after enough new input patterns, the old information is washed away. Adaptive resonance theory [ART] enables a network to be robust against arbitrary sequences of inputs in a self-regulated manner. It is self-organized in the sense that it is unsupervised.
