[Humanist] 24.581 linear problems in AI?

Humanist Discussion Group willard.mccarty at mccarty.org.uk
Mon Dec 13 07:26:54 CET 2010

                 Humanist Discussion Group, Vol. 24, No. 581.
         Centre for Computing in the Humanities, King's College London
                Submit to: humanist at lists.digitalhumanities.org

        Date: Sun, 12 Dec 2010 12:18:29 +0000
        From: Willard McCarty <willard.mccarty at mccarty.org.uk>
        Subject: linear problems and their solutions

Hiroaki Kitano, in "Challenges of evolvable systems: Analysis and future 
directions", LNCS 1259, p. 126, refers to "class-I" and "class-II" 
problems in computer science as having the following characteristics:

> *Discreteness*: Objects and features in the domain have a high degree
> of discreteness, allowing the domain to be mapped into a [sic] symbolic
> representations.
 > *Explicitness*: Rules governing the domain exist in an
> explicit form.
 > *Completeness*: A complete set of rules can be
> obtained.

Is this a standard characterisation of such problems? Kitano notes that 
in AI these problems can be solved by linear decomposition or linear 
approximation, i.e. that they are essentially linear problems. He notes,

> The basic assumptions of this approach are: experts knew [sic] necessary
> and sufficient knowledge for the task, and this expert knowledge can
> be expressed in symbolic form. It also assumes that the knowledge
> acquired is complete, correct, and consistent. Provided these
> assumptions hold, traditional AI techniques are a powerful means of
> problem solving.

Who has written best about problems in these terms and what their solutions require?

Many thanks.


Willard McCarty, Professor of Humanities Computing,
King's College London, staff.cch.kcl.ac.uk/~wmccarty/;
Professor, Centre for Cultural Research, University of Western Sydney,
Editor, Humanist, www.digitalhumanities.org/humanist;
Editor, Interdisciplinary Science Reviews, www.isr-journal.org.

More information about the Humanist mailing list