HOME   ::  Conference List   ::   Conference Paper
Hashimoto, T. and Ikegami, T. (1995) Communication Network of Symbolic Grammar Systems. In Y. Aizawa and et al., editors, Proceedings of the International Conference on Dynamical Systems and Chaos, pages 595--598. Singapore: World Scientific.
Bookmark:  

Full-text
   URL: http://www.jaist.ac.jp/~hash/papers/symgram-ICDC.ps.gz
   Cached: PDF-145K    PS-679K    PS.gz-62K   
   SAVE AS an easy-to-recall long filename:
      Filename format: author--year--title   PDF-145K    PS-62K    :: About GZip'd PS
      Filename format: author--year--title--journal|proceedings|...--pages   PDF-145K    PS-62K   

Related links
  Web search: Google Web Search   ::   Google Scholar
  Within this site: Cited by (1)    References (5)

Paper at a Glance 

Communication Network of
Symbolic Grammar Systems
TAKASHI HASHIMOTO 3 and TAKASHI IKEGAMI y
3 The Graduate School of Science and Technology, Kobe University
y Institute of Physics, College of Arts and Sciences, University of Tokyo
Abstract ­ Interacting agents with symbolic grammar are proposed in order to study the evolution of computational ability of agents. The algorithmic evolution of the formal grammar system is characterized by Chomsky's hierarchy 1 . Agents with a higher grammar can speak/recognize many more words than those with a lower one. However, when agents form a network, the higher Chomsky grammar is not always advantageous. It is shown that to speak/recognize commonly used words is more favorable in a network.
INTRODUCTION In this paper we present an evolutionary model of interacting agents with symbolic grammar. Our main concern is to see how a higher level of grammar evolves from the lower ones. Regarding computational ability as a measure of evolution, we naturally ask, what evolutionary dynamics can elaborate computational ability? Introducing an ensemble of agents, we discuss its evolution. We see evolutionary pathways of climbing up Chomsky's hierarchy and show how they are avoided by the ensemble. DYNAMICAL MODELING OF COMMUNICATION NETWORK We characterize each agents by a set of rewriting rules, V !a, where V is a non­terminal symbol and a is a list of non­terminal and terminal symbols. To generate sentences, the agents apply their rules from the left to right hand side. If there are several rewriting paths, one of them is selected randomly. To recognize sentences, rules are applied in the opposite direction. If a agent rewrites a sentence to the start symbol ''S'' in finite rewriting steps, the agent can recognize it. When we make a network of agents, each agent speaks in turn by using its grammar and tries to recognize the sentences spoken by the others. 3 e­mail: toshiwo@sacral.c.u­tokyo.ac.jp y e­mail: ikeg@sacral.c.u­tokyo.ac.jp 1 ...
BibTex
@inproceedings{hashimoto95communicationNetwork,
  author={T. Hashimoto and T. Ikegami},
  title={Communication Network of Symbolic Grammar Systems},
  year={1995},
  pages={595--598},
  address={Singapore},
  editor={Y. Aizawa and et al.},
  publisher={World Scientific},
  booktitle={Proceedings of the International Conference on Dynamical Systems and Chaos},
  url={http://www.isrl.uiuc.edu/~amag/langev/paper/hashimoto95communicationNetwork.html}
}

 HOME   ::  Conference List   ::   Conference Paper Comments to: junwang4 you-know-at gmail.com Last update: 2/2/08