Semiotic Oriented Autonomous Intelligent Systems Software Engineering

1
Semiotic Oriented Autonomous Intelligent Systems
Software Engineering

• State University of Campinas
• UNICAMP - Brazil

Rodrigo Gonçalves Ricardo Gudwin
2
Intelligent Systems Evolution

• Increasing complexity and autonomy

Complexity
Problems (use cases, tool oriented)
Engineering methods
3
Tool oriented x model oriented

• Tool oriented
• Model oriented
• Analysis Project Design

NNet
GA
ESPECIFICATION This work introduces a first
proposal on how to use semiotics in order to
improve software engineering methods, when
intelligent autonomous systems are targeted.
First we investigate the current flaws in
software engineering, concerning intelligent
autonomous systems. Then we propose a knowledge
taxonomy, based on semiotic ideas, aiming a tool
to
?
4
Semiotics software engineering

• Semiotics tool to understand knowledge
  processing and communication
• Semiotics can be used in analysis phase of any
  software engineering method

ESPECIFICATION This work introduces a first
proposal on how to use semiotics in order to
improve software engineering methods, when
intelligent autonomous systems are targeted.
First we investigate the current flaws in
software engineering,
Information structure flow
Architectures
Information structure flow
Information structure flow
Project Design
Semiotic analysis
5
SemioticsBasic concepts

• Knowledge unit
• Umwelt
• Semiosis

6
SemioticsKnowledge taxonomy

• Knowledge taxonomy
• Functionality
• Structure
• Both
• Structure
• Rematic R
• Dicent D
• Functionality
• Designative
• Apraisive
• Prescriptive
• Both (special case)
• Argumentative

RIcObSp
RIcSeG
Sensors
RIcObG
RIn
RSy
DSy
DIc
RIcSeSp
Actuator
7
Example Artificial Fish

• Terzopoulos, D., Artificial Fishes, Artificial
  Life, Volume 1, Number 4, MIT Press, 1994
• Terzopoulos approach
• It would not be possible if the system were more
  complex

ESPECIFICATION This work introduces a first
proposal on how to use semiotics in order to
improve software engineering methods, when
intelligent autonomous systems are targeted.
First we investigate the current flaws in
software engineering, concerning intelligent
autonomous systems. Then we propose a knowledge
taxonomy, based on semiotic ideas, aiming a tool
to
8
ExampleArtificial Fish

• Proposed approach

Architectures
ESPECIFICATION This work introduces a first
proposal on how to use semiotics in order to
improve software engineering methods, when
intelligent autonomous systems are targeted.
First we investigate the current flaws in
software engineering, concerning intelligent
autonomous systems. Then we propose a knowledge
taxonomy, based on semiotic ideas, aiming a tool
to
Semiotic analysis and project
Information domain model
9
Future works

• Work in progress
• Future works
• Study of other architectures than GSM
• Specification of a framework to help in the
  design phase. Based on (fielded) object networks.
• How these concepts apply in a specific method
  such as UML (UML extensions)
• Implementation of an example

10
Conclusions

• IS complexity and autonomy is increasing bringing
  the necessity of software engineering methods
• Traditional software engineering methods does not
  conveniently apply to intelligent systems.
• Semiotics can be used as a tool to understand the
  information domain of a problem
• Semiotics can be used in the analysis phase of
  any software engineering method.
• Next works will introduce how these concepts
  apply in a specific method such UML (Unified
  Modeling Language)