A Software Agent with a Self? Machine Consciousness: Complexity Aspects 29 Sept to 1st October...

Preview:

Citation preview

A Software Agent with a Self?A Software Agent with a Self?

Machine Consciousness: Complexity Aspects 29 Sept to 1st October '03,

ISI, Torino, Italy

Stan Franklinand the

“Conscious” Software Research Group

September 30, 2003 Stan Franklin—Machine Consciousness—Torino

2

IDA: an Intelligent Distribution Agent

IDA: 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

September 30, 2003 Stan Franklin—Machine Consciousness—Torino

3

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

September 30, 2003 Stan Franklin—Machine Consciousness—Torino

4

Contexts at work

September 30, 2003 Stan Franklin—Machine Consciousness—Torino

5

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

September 30, 2003 Stan Franklin—Machine Consciousness—Torino

7

The Various SelvesThe Various Selves

September 30, 2003 Stan Franklin—Machine Consciousness—Torino

8

Baars—Self as the enduring contextBaars—Self as the enduring context

September 30, 2003 Stan Franklin—Machine Consciousness—Torino

9

Baar’s Self-SystemBaar’s Self-System

• Higher Levels of the dominant context hierarchy• Overarching • Enduring• Includes both goal contexts

and conceptual contexts• Not available to introspection• Includes self-concept • Executive function (volitional self)

September 30, 2003 Stan Franklin—Machine Consciousness—Torino

10

Damasio’s Proto-SelfDamasio’s Proto-Self

• Moment to moment

• Unconscious

• Tracks state of

organism

• Proprioception

• Homeostasis

• Implement in IDA

• Operating system parameters

• Behavior net parameters

• Memory (SDM) parameters

• Number of threads

• Computer memory usage

September 30, 2003 Stan Franklin—Machine Consciousness—Torino

11

Minimal (Core) SelfMinimal (Core) Self

• Gallagher, Damasio, Donald, many others

• Self as subject—self-as-subject fringe codelets

• Self as experiencer— fringe codelets (?), always present (?),

function (?)

• Self as agent— intention codelets (are generated when one creates an intention to look for opportunities to do something).

September 30, 2003 Stan Franklin—Machine Consciousness—Torino

12

Extended SelfExtended Self

• Autobiographical Self—Hierarchical autobiographical memory (SDM?)

• Self-concept (Blackmore—Selfplex)

• Self-beliefs—Semantic Memory (SDM?)

• Intentions—Intention Codelets

• Volitional Self (Baars—Executive Function)

• Ideomotor theory as a behavior stream

• Intention codelet for each volitional goal (intention)

• Narrative Self (Gazzaniga—Interpreter)

• Understanding report requests and making reports

September 30, 2003 Stan Franklin—Machine Consciousness—Torino

13

Understanding Requests for a Report

Understanding Requests for a Report

• Perception codelets

• Slipnet nodes and links

• Working memory tags

September 30, 2003 Stan Franklin—Machine Consciousness—Torino

14

Making ReportsMaking Reports

• Motivation—pleasure at responding to queries about self

• Conscious of request—request attention codelets

• Report contents of consciousness • Report behavior streams• Report message scripts

September 30, 2003 Stan Franklin—Machine Consciousness—Torino

15

Is IDA a conscious artifact?Is IDA a conscious artifact?

• Functional vs. Phenomenal Consciousness

• IDA is functionally conscious

• I doubt IDA is phenomenally conscious

• No good argument to support that doubt

• Can you give me one?

• Would IDA with a self be phenomenally conscious?

September 30, 2003 Stan Franklin—Machine Consciousness—Torino

16

PredictionsPredictions

SERIOUSWe humans will build “conscious”software agentsand/or “conscious” robots,so intelligent, so sophisticated, and so communicative that people will simply

assume that they are sentient.

SPECULATIVE (Sloman)Some day, these same “conscious” software agentswill find themselves discussingwhether or not humans canreally experience qualia.

September 30, 2003 Stan Franklin—Machine Consciousness—Torino

17

Web and Email AddressesWeb and Email Addresses

• Stan Franklin • franklin@memphis.edu—• www.cs.memphis.edu/~franklin

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

Recommended