A Computational Semiotics Approach for Soft Computing

1
A Computational Semiotics Approach for Soft
Computing

• Ricardo R. Gudwin
• Fernando A.C. Gomide
• DCA-FEEC-UNICAMP

2
Introduction

• Computational Intelligence and Soft Computing
• model intelligent behavior using ideas from
  biology and the definition and use of uncertainty
• fuzzy systems
• neural networks
• evolutive systems
• Hybrid Models
• neuro-fuzzy
• neuro-genetic
• fuzzy-genetic

3
Introduction

• Computational Semiotics
• Emulation of the process of Semiosis in a
  computer system
• Mathematically define concepts from semiotics in
  order to be used in a computer system
• Object (agent)-oriented structure
• Meta-theoretical tool designed to formalize
  intelligent systems
• Unify the representations used to formalize the
  different behaviors found within soft computing

4
Fundamental Transformations

• Argumentative knowledge
• arguments
• knowledge of transforming knowledge
• Three main arguments
• knowledge extraction (deduction)
• knowledge generation (induction)
• knowledge selection (abduction)
• Selection and Internal Functions in an active
  object
• Building blocks for intelligent systems (soft
  computing)

5
Conclusions

• Computational Semiotics
• aiming at an unified formal model for soft
  computing
• extending soft computing through hybrid systems
• focus on the knowledge process embedded in each
  soft computing technique (fuzzy, neural, genetic)
• Use of deductive, inductive and abductive
  arguments to build intelligent behavior
• Formal model easily converted into a
  computational algorithm
• General enough to accommodate specific details of
  each soft computing technique
• Do not compete with the current developments for
  each technique