What is a Knowledge Representation?

Dec 06, 2010 03:16 PM

Excellent paper by Randy Davis, Howard Shrobe, and Pete Szolowits

R. Davis, H. Shrobe, and P. Szolovits. What is a Knowledge Representation? <cite>AI Magazine</cite>, 14(1):17-33, 1993

<ul>

<li>A knowledge representation (KR) is most fundamentally a <code>surrogate</code>, a substitute for the thing itself, used to enable an entity to determine consequences by thinking rather than acting, i.e., by reasoning about the world rather than taking action in it.</li>

<li>It is a <code>set of ontological commitments</code>, i.e., an answer to the question: In what terms should I think about the world?</li>

<li>It is a <code>fragmentary theory of intelligent reasoning</code>, expressed in terms of three components: (i) the representation's fundamental conception of intelligent reasoning; (ii) the set of inferences the representation <code>sanctions</code>; and (iii) the set of inferences it <code>recommends</code>.</li>

<li>It is a <code>medium for pragmatically efficient computation</code>, i.e., the computational environment in which thinking is accomplished. One contribution to this pragmatic efficiency is supplied by the guidance a representation provides for organizing information so as to facilitate making the recommended inferences.</li>

<li>It is a <code>medium of human expression</code>, i.e., a language in which we say things about the world.</li>

</ul>

it makes the case that one significant part of the representation endeavor--capturing and representing the richness of the natural world--is receiving insufficient attention.

Role 1:  KR is a Surrogate:

reasoning is internal;  most things the agent wants to reason about are external, eg bicycle assembly.

KR is a stand-in for the things that exist in the world.

operations on and with reprs substitute for operations on the real thing.

In this view, reasoning itself is in a part a substitute for action in the world.

there must be some form of correspondence btwn the surrogate and the its referent:  that is its semantics.

all our reps are inaccurate: they all simplify.

reps are also surrogates for abstract notions like actions, processes, beliefs, causality, categories.

Imperfect surrogates mean incorrect inferences are inevitable

Two important consequences follow from the inevitability of imperfect surrogates. One consequence is that in describing the natural world, we must inevitably lie, by omission at least. At a minimum we must omit some of the effectively limitless complexity of the natural world; our descriptions may in addition introduce artifacts not present in the world.

All broad based reasoning about the natural world MUST reach some conclusions that are not correct.

it makes sense to evaluate pragmatically the relative costs and benefits that come from using both sound and unsound reasoning methods

ROLE II:  A KR is a set of ONtological Commitments:  each KR attends to some things and ignores others:

so we are making a deicison about how and what to see in the world.

The commitments are in effect a strong pair of glasses bringing some parts of the world into sharp focus at the expense of

blurring others.

judicious selection of commitments focuses attn on aspects we believe to be relevant.

The focusing effect is an essential part of what a representation offers, because the complexity of the natural world is overwhelming. We (and our reasoning machines) need guidance in deciding what in the world to attend to and what to ignore. The glasses supplied by a representation can provide that guidance: In telling us what and how to see, they allow us to cope with what would otherwise be untenable complexity and detail. Hence the ontological commitment made by a representation can be one of the most important contributions it offers.

What is a Knowledge Representation?

Attachments