background to the ida model

23
Background to the IDA Model Stan Franklin and the ‘Conscious’ Software Research Group Machine Consciousness Workshop, Torino, Fall 2003

Upload: blake

Post on 19-Jan-2016

24 views

Category:

Documents


0 download

DESCRIPTION

Background to the IDA Model. Stan Franklin and the ‘Conscious’ Software Research Group Machine Consciousness Workshop, Torino, Fall 2003. I. A. D. Interne t. IDA: an Intelligent Distribution Agent. Dialogue with sailors Read personnel data Check job requisition lists - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Background to the IDA Model

Background to the IDA Model

Background to the IDA Model

Stan Franklinand the

‘Conscious’ Software Research Group

Machine Consciousness Workshop, Torino, Fall 2003

Page 2: Background to the IDA Model

September 29, 2003 Stan Franklin—Machine Consciousness—Torino

2

IDA: an Intelligent Distribution AgentIDA: an Intelligent Distribution Agent

Detailer

Telephone

Dialogue with sailors

Read personnel data

Check job requisition lists

Enforce Navy policies

Choose jobs to offer members

Negotiate with them about jobsInternet

IDA

Page 3: Background to the IDA Model

September 29, 2003 Stan Franklin—Machine Consciousness—Torino

3

IDA: a ‘conscious’ software agentIDA: a ‘conscious’ software agent

Autonomous Agent GW Theory

Computational ModelConceptual Model

Page 4: Background to the IDA Model

September 29, 2003 Stan Franklin—Machine Consciousness—Torino

4

Global Workspace Theory

• A psychological theory of consciousness

• The nervous system is a distributed parallel system with many different specialized processors

• Global workspace contains a coalition of processors

• Broadcasts globally to all other processors

• Recruit other processors needed for any degree of novel or problematic situation

• Explains limited capacity and seriality

Page 5: Background to the IDA Model

September 29, 2003 Stan Franklin—Machine Consciousness—Torino

5

Contexts at work

Page 6: Background to the IDA Model

September 29, 2003 Stan Franklin—Machine Consciousness—Torino

6

Why a ‘Conscious’ Agent?Why a ‘Conscious’ Agent?

• Flesh out the theory with detailed architecture and mechanisms

• Hypotheses for cognitive scientists and neuroscientists

• Produce flexible, adaptive, human-like software

• Want smart agents? Model them after humans.

Page 7: Background to the IDA Model

September 29, 2003 Stan Franklin—Machine Consciousness—Torino

7

Modules and MechanismsModules and Mechanisms

• Perception—Copycat Architecture—Hofstadter

• Action Selection—Behavior Net—Maes

• Episodic Memory—Sparse Distributed Memory—Kanerva

• Emotions—Pandemonium Theory—Jackson

• Metacognition—Fuzzy Classifier Systems—Holland

• Learning—Copycat Architecture, Reinforcement

• Constraint Satisfaction—Linear Functional

• Language Generation—Pandemonium Theory

• Deliberation—Pandemonium Theory

• ‘Consciousness’ —Pandemonium Theory

Page 8: Background to the IDA Model

September 29, 2003 Stan Franklin—Machine Consciousness—Torino

8

IDA’s ArchitectureIDA’s Architecture

‘Consciousness’Perception

Metacognition

WorkingMemory

EpisodicMemory

Behavior Net

Emotions

DatabasePerception

ConstraintSatisfaction

Deliberation NegotiationProblemSolving

Expectation&

Automization

Page 9: Background to the IDA Model

September 29, 2003 Stan Franklin—Machine Consciousness—Torino

9

Processing in IDAProcessing in IDA

A continuing iteration of a cognitive cycle of activities involving:

• Perception• Working memory• Transient episodic memory• Long-term declarative memory• ‘Consciousness’• Action selection• Motor activity

Page 10: Background to the IDA Model
Page 11: Background to the IDA Model

September 29, 2003 Stan Franklin—Machine Consciousness—Torino

11

Cognitive Cycle ProcessingCognitive Cycle Processing

• Hypothesis— Like IDA’s, human cognitive processing is via a continuing sequence of Cognitive Cycles

• Duration— Each cognitive cycle takes roughly 200 ms with steps 1 through 5 occupying about 80 ms

• Overlapping— Several cycles may have parts running simultaneously in parallel

• Seriality— Consciousness maintains serial order and the illusion of continuity

• Start— Cycle may start with action selection instead of perception

Page 12: Background to the IDA Model

September 29, 2003 Stan Franklin—Machine Consciousness—Torino

12

Levels of abstractionLevels of abstraction

• High level• behaviors• message type nodes• emotions• metacognitive actions• etc.

• Low level• codelets

Page 13: Background to the IDA Model

September 29, 2003 Stan Franklin—Machine Consciousness—Torino

13

CodeletsCodelets

• Small pieces of code each performing a simple, specialized task

• Many acts as demons, watching for a chance to act• Most subserve some high level entity, e.g.

• behavior• slipnet node• metacognitive action

• Some codelets work on their own, e.g.• watching for incoming mail• checking for time and place conflicts

• Codelets do almost all the work• IDA is a multi-agent system

Page 14: Background to the IDA Model

September 29, 2003 Stan Franklin—Machine Consciousness—Torino

14

Perception via a Slipnet

preference acceptance information request

location

San Diego Miami Norfolk Jacksonville. . .

Norfolk norfolk nor NRFK

Page 15: Background to the IDA Model

September 29, 2003 Stan Franklin—Machine Consciousness—Torino

15

Coalitions and Consciousness

• Coalition manager

• Spotlight manager

• Broadcast mechanism

Page 16: Background to the IDA Model

September 29, 2003 Stan Franklin—Machine Consciousness—Torino

16

A Behavior StreamA Behavior Stream

Find and movea template

Compose an acknowledgment

Find an emailaddress

Drive to Acknowledge

From the Sidelines

Activation from drive

Activation from the environment, external or internal

Send an acknowledgement

Page 17: Background to the IDA Model

September 29, 2003 Stan Franklin—Machine Consciousness—Torino

17

Behavior Net in ActionBehavior Net in Action

Behavior net templates

Behavior net

Side lines

Playing field

StandsWorkingMemory

Broadcast

Page 18: Background to the IDA Model

September 29, 2003 Stan Franklin—Machine Consciousness—Torino

18

‘Consciousness’ in Action‘Consciousness’ in Action

Associative Memory

Working memory

Focus

PlayingField

Stands

OutgoingMessage

JobList

Page 19: Background to the IDA Model

September 29, 2003 Stan Franklin—Machine Consciousness—Torino

19

DeliberationDeliberation

• Faced with a goal or problem • Imagine possible plans or solutions

• Scenarios• Routes • Internal virtual reality—Dawkins

• Evaluate them• Using reason• Using emotions

• Choose among them

Page 20: Background to the IDA Model

September 29, 2003 Stan Franklin—Machine Consciousness—Torino

20

Deliberation in ActionDeliberation in Action

Associative Memory

Working memory

Focus

PlayingField

Stands

JobList

DetachDate

DetachDate

DetachDate

LeaveTime

LeaveTime

Leave Time

Page 21: Background to the IDA Model

September 29, 2003 Stan Franklin—Machine Consciousness—Torino

21

Ideomotor TheoryIdeomotor TheoryWilliam James (circa 1890) ----- Bernard Baars (1988)

Voluntary vs non-voluntary action

Theory of voluntary action• Proposers—propose a course of action• Objectors—raise objections to such a course of action• Supporters—lend support to such a course of action• Auctioneer—wields the gavel

Page 22: Background to the IDA Model

September 29, 2003 Stan Franklin—Machine Consciousness—Torino

22

Ideomotor Theory in ActionIdeomotor Theory in Action

Idea pops to mind (proposer)—no objection (objector)—do itObjection (objector)—don’t do itObjection then support (supporter)—do it

Different proposal—no objection—do itDifferent proposal—original proposal— no objection—do it

Last unopposed proposal is acted upon

Page 23: Background to the IDA Model

September 29, 2003 Stan Franklin—Machine Consciousness—Torino

23

Web and Email AddressesWeb and Email Addresses

• Stan Franklin • [email protected]• www.cs.memphis.edu/~franklin

• ‘Conscious’ Software Research Group• www.cs.memphis.edu/~csrg