A Strategy for Efficient Persuasion Dialogues - Agents and Artificial Intelligence - page 326

A Strategy for Efficient Persuasion Dialogues

Katie Atkinson 1 , Priscilla Bench-Capon 2 , and Trevor Bench-Capon 1

1 Department of Computer Science, University of Liverpool, U.K.
2 Open University, Milton Keynes, U.K.
{ katie,tbc } @liverpool.ac.uk, [e-mail]

Abstract. Several dialogue types, including inquiry, persuasion and delibera-
tion, transfer information between agents so that their beliefs and opinions may
be revised. The speech acts in different dialogue types have different pragmatic
implications. For a representative sub-type of persuasion dialogues we consider
how they can be conducted efficiently, in terms of minimising the expected trans-
fer of information, and develop a strategy for efficient persuasion by exploiting
the pragmatic implications. We demonstrate that our strategy is optimal for this
sub-type.

1 Introduction

Several authors have developed dialogue protocols for use in agent systems based on
the typology of Walton and Krabbe [9], including [2] for inquiry dialogues; [6] for per-
suasion; and [5] for deliberation. It is argued in [9] that, unlesss agents recognise (and
agree on) the type of dialogue they are engaged in, misunderstandings arise, fallacies
become possible and the conversation may break down. Even so, the distinctions have
rarely been made precise, and confusion is increased because inquiry, persuasion and
deliberation all make use of a similar set of speech acts. In this paper we will give a
precise characterisation of the distinctive features of a sub-type of persuasion dialogues
and distinguish them from deliberation and inquiry dialogues. Our analysis will draw
attention to the pragmatic meaning of utterances which vary across dialogue types and
give them their distinctive character. We offer a clear specification of these persuasion
dialogues, and an optimal strategy for conducting a dialogue of this representative sub-
type. Our contribution is thus the analysis yielding the distinguishing features of per-
suasion dialogues, and an optimal strategy for a common class of persuasion dialogues.
One common situation that gives rise to both persuasion and deliberation is when
an agent must choose between several options, each of which has several features that
can serve as reasons for and against the choice, and to which agents will ascribe dif-
ferent degrees of importance depending on their personal preferences. When buying a
car one person will be most interested in safety, another in speed, another in comfort.
For cameras, some will require the best quality, whereas others will value more highly
compactness or a low price. In such situations an agent will often need to deal with a
series of salespeople, each trying to overturn the agent's currently preferred option, or
to consider recommendations from other agents. A very common example in AI con-
ference papers is choosing a restaurant for lunch or an evening dinner. Typically the
individual agents will have only incomplete and often ill-founded or outdated beliefs