Alan K. Mackworth's Publications

Sorted by DateClassified by Publication TypeSorted by First Author Last NameClassified by Author Last Name

Formal Specification of Performance Metrics for Intelligent Systems

Ying Zhang and Alan K. Mackworth. Formal Specification of Performance Metrics for Intelligent Systems. In Proceedings of the Workshop on Performance Metrics for Intelligent Systems, pp. 33–34, National Institute of Standards and Technology, Washington, DC, August 2000.

Download

[PDF]38.6kB  

Abstract

There are now so many architectures for intelligent systems: deliberative planning vs. reactive acting, behavioral subsuming vs. hierarchical structuring, machine learning vs. logic reasoning, and symbolic representation vs. procedural knowledge. The arguments from all schools are all based on how natural systems (i.e., biologically inspired, from basic forms of life to high level intelligence) work by taking the parts that support their architectures. In this paper, we take an engineering point of view, i.e., by using requirements specification and system verification as the measurement tool. Since most intelligent systems are realtime dynamic systems (all lives are), requirements specification should be able to represent timed properties. We have developed timed "-automata that fit to this purpose. We will present this formal specification, examples for specifying requirements and a general procedure for verification.

BibTeX

@InProceedings{WPMIS00,
  author =	 {Ying Zhang and Alan K. Mackworth},
  title =	 {Formal Specification of Performance Metrics for Intelligent Systems},
  year =	 {2000}, 
  month =        {August},
  booktitle =	 {Proceedings of the Workshop on Performance Metrics for Intelligent Systems},
  address =      {National Institute of Standards and Technology, Washington, DC},
  pages =         {33--34},
  abstract =	 { There are now so many architectures for intelligent systems:
                   deliberative planning vs. reactive acting, behavioral subsuming
                   vs. hierarchical structuring, machine learning vs. logic reasoning,
                   and symbolic representation vs. procedural knowledge. The
                   arguments from all schools are all based on how natural systems
                   (i.e., biologically inspired, from basic forms of life to high level
                   intelligence) work by taking the parts that support their
                   architectures. In this paper, we take an engineering point of view,
                   i.e., by using requirements specification and system verification
                   as the measurement tool. Since most intelligent systems are realtime
                   dynamic systems (all lives are), requirements specification
                   should be able to represent timed properties. We have developed
                   timed "-automata that fit to this purpose. We will present this
                   formal specification, examples for specifying requirements and a
                   general procedure for verification.},
  bib2html_pubtype ={Refereed Conference Proceeding},
  bib2html_rescat ={},
}

Generated by bib2html.pl (written by Patrick Riley ) on Wed Apr 23, 2014 19:08:34