Information Technology Reference
In-Depth Information
Essentially an agent starts with incomplete information. In a dialogue agents can
pool information that can extend their knowledge and perhaps revise their beliefs. At
any time an agent has a belief about the utility of various options, but this is tentative
because of the incompleteness and potential incorrectness of its beliefs. The purpose
of persuasion is thus to acquire information and verify the current beliefs to confirm
the current preference or establish a new one. In a deliberation not only is information
about facts exchanged, but also information about the importance of particular attributes
for particular agents in determining the utility of an option. Thus an inquiry concerns
a particular item of information. A persuasion concerns information relating to the set
of attributes relevant to determining a particular preference. A deliberation concerns
this information, and information as to what attributes might be used to determine this
preference. Persuasion and deliberation can thus be seen as composed of a series of
inquiry dialogues, with the context determining the subjects of inquiry.
2.1 Definitions
An example, which illustrates our notation, is given in Section 3. The reader might
find it helpful to refer to this in conjunction with the following definitions for concrete
examples of their use.
The knowledge bases of agents can be partitioned into
factual elements
,onwhich
agents should agree, used when the direction of fit is from world to beliefs, and
prefer-
ence elements
, which represent their own individual desires, tastes and aspirations, and
are used when the direction of fit is from desires to the world. Thus, the preference ele-
ments represent the way the agent evaluates possible options to determine which wishes
to bring about, and how it evaluates objects and situations for value judgements such as
best car
and
acceptable restaurant
.
Definition 1.
Let
AG
denote a set of agents, each of which,
Ag
AG
, has a knowledge
base
KB
Ag
.
KB
Ag
is partitioned into factual elements denoted by
KB
Ag
F
∈
and prefer-
ence information denoted by
KB
Ag
P
.
KB
Ag
F
comprises facts, strict rules and defeasible
rules.
KB
A
P
comprises rules to determine the utility for
Ag
of certain items based on
their attributes, and the weights used by these rules. These preference elements are
defined below.
Agents expand their
KB
F
by taking information from one another, but
KB
P
remains
fixed throughout the dialogue. Whereas, because it is intended to fit the world,
KB
F
is objective,
KB
P
represents the personal preferences of an individual agent, and so is
entirely local to the agent concerned. We will use
f
for factual propositions, and
p
Ag
(to be read as “
p
is the case for
Ag
”) for propositions based on preferences. We will not
represent actions separately, so that
p
Ag
may represent either propositions such as
Roux
Brothers is an acceptable restaurant for Ag
or propositions such as
it is desirable for
Ag
A
that
Ag
B
φ
.
Definition 2.
Let
PROP
be a set of atomic propositions and let
f
PROP
be a
factual proposition
and
p
Ag
∈ PROP
be a proposition based on preferences. Let
KB
Ag
F
∈
f
denote that a factual proposition can be defeasibly shown from
KB
A
F
,us-
ing an appropriate reasoning engine (e.g. [2] uses Defeasible Logic [3]), without using
