Abstract for AISB 2000: How to Design a Functioning Mind
AUTHOR: Brian Logan, School of Computer Science & IT
University of Nottingham
TITLE: A design study for an AFP architecture
ABSTRACT:
This working paper describes a design study for a variant of the
`Attention Filter Penetration' (AFP) three layer architecture. In this
architecture, the activities of the agent are distributed across three
concurrently executing layers: a reactive layer in which detection of
internal or external conditions immediately generates new internal or
external response, a deliberative layer responsible for `what if'
reasoning and planning capabilities, and a meta-management layer which
provides self monitoring, self evaluation and self-redirection including
control of attention. An attention filter with dynamically varying
interrupt threshold protects the resource-limited deliberative and
meta-management layers when dealing with tasks that are important,
urgent and resource consuming.
The aim is to design an architecture for a system which can control its
efforts to achieve one or more goals in some domain, and for the system
to be able to distinguish when it is in control of its own processing
and when it isn't.
The design study is based on two main principles: that deliberation (and
meta-deliberation or management) is the technique of last resort for an
agent, and is only used when no reactive behaviour is clearly applicable
in a given situation; and deliberation (and management) is just
something that some reactive systems do. We therefore focus first on
the reactive layer of the architecture to clarify what it can and can't
do, and to outline how it does what it does, before attempting to show
how the deliberative and management layers can be implemented as
particular kinds of reactive behaviour. In doing so, the aim is not an
explanatory reduction of the deliberative or meta-management layers to
the reactive layer, rather the aim is to sketch an implementation of the
virtual machines which operate at these layers in terms of the
primitives available at the reactive layer. Of particular interest
therefore is the interaction between the reactive and deliberative
layers: both the generation of new motives or goals by the reactive
layer and when and how these are scheduled for processing at the
deliberative layer, and the controlled execution or modulation of
reactive behaviours by the deliberative layer.
The paper attempts to identify those parts of the design which seem
fairly uncontroversial, and to highlight those issues which have yet to
be resolved. In several cases, a number of possible approaches to an
unresolved issue are identified. Such speculations are not intended to
exhaustive enumerations of the options, rather they attempt to indicate
the current state of work on a topic and illustrate insofar as this is
possible at this stage, some of the main issues that would have to be
addressed by any solution.
=======================================================================
SHORT CV
I am interested in architectures for autonomous agents, particularly
human-like agents.
I am currently working with Aaron Sloman on a project Evolvable virtual
information processing architectures for human-like minds. The aim of
the project is to explore information processing models of mind by
combining philosophical conceptual analysis and philosophically inspired
top-down design, along with bottom-up use of Artificial Intelligence
(AI) techniques to produce software and demonstrations. We also attempt
to take account of ideas and empirical results from other relevant
disciplines, including brain science, psychology, and theoretical
biology.
I am also interested in tools for developing agents, and (with Jeremy
Baxter) co-chaired a workshop on Software Tools for Developing Agents,
held at the 15th National Conference on Artificial Intelligence
(AAAI-98).
Recent publications:
Building Cognitively Rich Agents Using the SIM_AGENT Toolkit (with A.
Sloman), Communications of the ACM, March 1999/Vol. 42, No. 3, pp 71-77.
A Framework for the Distributed Simulation of Agent-based Systems (with
G. Theodoropoulos), in Proceedings of the 13th European Simulation
Multiconference (ESM'99), 1999, pp 58-65.
Software Tools for Developing Agents: Papers from the 1998 Workshop
(with J. Baxter) eds., Technical Report WS-98-10, AAAI Press, Menlo
Park, CA, 1998.
Architectures and Tools for Human-Like Agents (with A. Sloman), in
Proceedings of the Second European Conference on Cognitive Modelling,
1st-4th April 1998, University of Nottingham, Nottingham, pp 58--65.