analogy making n.
Download
Skip this Video
Loading SlideShow in 5 Seconds..
Analogy-Making PowerPoint Presentation
Download Presentation
Analogy-Making

Loading in 2 Seconds...

play fullscreen
1 / 148

Analogy-Making - PowerPoint PPT Presentation


  • 115 Views
  • Uploaded on

Analogy-Making. Consider the following cognitive activities. Recognition:. Recognition: A child learns to recognize cats and dogs in books as well as in real life. Recognition: A child learns to recognize cats and dogs in books as well as in real life.

loader
I am the owner, or an agent authorized to act on behalf of the owner, of the copyrighted work described.
capcha
Download Presentation

Analogy-Making


An Image/Link below is provided (as is) to download presentation

Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author.While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server.


- - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - -
Presentation Transcript
slide4
Recognition:
    • A child learns to recognize cats and dogs in books as well as in real life.
slide5
Recognition:
    • A child learns to recognize cats and dogs in books as well as in real life.
slide6
People can recognize letters of the alphabet, e.g., ‘A’, in many different typefaces and handwriting styles.
slide7
People can recognize letters of the alphabet, e.g., ‘A’, in many different typefaces and handwriting styles.
slide9
People can recognize styles of music:
    • “That sounds like Mozart”
slide10
People can recognize styles of music:
    • “That sounds like Mozart”
    • “That’s a muzak version of ‘Hey Jude’”
slide11
People can recognize styles of music:
    • “That sounds like Mozart”
    • “That’s a muzak version of ‘Hey Jude’”
  • People can recognize abstract situations:
slide12
People can recognize styles of music:
    • “That sounds like Mozart”
    • “That’s a muzak version of ‘Hey Jude’”
  • People can recognize abstract situations:
    • A “Cinderella story”
    • “Another Vietnam”
    • “Monica-gate”
    • “Shop-aholic”
slide14
People make scientific analogies:
    • “Biological competition is like economic competition” (Darwin)
slide15
People make scientific analogies:
    • “Biological competition is like economic competition” (Darwin)
    • “The nuclear force is like the electromagnetic force” (Yukawa)
slide16
People make scientific analogies:
    • “Biological competition is like economic competition” (Darwin)
    • “The nuclear force is like the electromagnetic force” (Yukawa)
    • “The computer is like the brain” (von Neumann)
slide17
People make scientific analogies:
    • “Biological competition is like economic competition” (Darwin)
    • “The nuclear force is like the electromagnetic force” (Yukawa)
    • “The computer is like the brain” (von Neumann)
    • “The brain is like the computer” (Simon, Newell, etc.)
slide19
People make unconscious analogies

Man:“I’m going shopping for a valentine for my wife.”

slide20
People make unconscious analogies

Man:“I’m going shopping for a valentine for my wife.”

Female colleague:“I did that yesterday.”

slide21
People make unconscious analogies

Newly married woman:“I often forget my new last name”

slide22
People make unconscious analogies

Newly married woman:“I often forget my new last name”

Man: “I have that trouble every January”

slide23
People make unconscious analogies

Computer scientist: “I’m in artificial intelligence because it’s a mixture of psychology, philosophy, linguistics, and computer science”

slide24
People make unconscious analogies

Computer scientist: “I’m in artificial intelligence because it’s a mixture of psychology, philosophy, linguistics, and computer science”

Architect: “That’s the reason I’m in architecture”

idealizing analogy making
Idealizing analogy-making

abc ---> abd

ijk--->

idealizing analogy making1
Idealizing analogy-making

abc ---> abd

ijk ---> ijl(replace rightmost

letter by successor)

idealizing analogy making2
Idealizing analogy-making

abc ---> abd

ijk ---> ijl(replace rightmost

letter by successor)

ijd(replace rightmost

letter by ‘d’)

idealizing analogy making3
Idealizing analogy-making

abc ---> abd

ijk ---> ijl(replace rightmost

letter by successor)

ijd(replace rightmost

letter by ‘d’)

ijk(replace all

‘c’sby ‘d’s)

idealizing analogy making4
Idealizing analogy-making

abc ---> abd

ijk ---> ijl(replace rightmost

letter by successor)

ijd(replace rightmost

letter by ‘d’)

ijk(replace all

‘c’sby ‘d’s)

abd(replace any

string by ‘abd’)

idealizing analogy making5
Idealizing analogy-making

abc ---> abd

iijjkk ---> ?

idealizing analogy making6
Idealizing analogy-making

abc ---> abd

iijjkk ---> iijjkl

Replace rightmost letter by successor

idealizing analogy making7
Idealizing analogy-making

abc ---> abd

iijjkk ---> ?

idealizing analogy making8
Idealizing analogy-making

abc ---> abd

iijjkk ---> iijjll

Replace rightmost “letter” by successor

idealizing analogy making9
Idealizing analogy-making

abc ---> abd

kji ---> ?

idealizing analogy making10
Idealizing analogy-making

abc ---> abd

kji ---> kjj

Replace rightmost letter by successor

idealizing analogy making11
Idealizing analogy-making

abc ---> abd

kji ---> ?

idealizing analogy making12
Idealizing analogy-making

abc ---> abd

kji ---> lji

Replace “rightmost” letter by successor

idealizing analogy making13
Idealizing analogy-making

abc ---> abd

kji ---> ?

idealizing analogy making14
Idealizing analogy-making

abc ---> abd

kji ---> ?

idealizing analogy making15
Idealizing analogy-making

abc ---> abd

kji ---> kjh

Replace rightmost letter by “successor”

idealizing analogy making16
Idealizing analogy-making

abc ---> abd

mrrjjj ---> ?

idealizing analogy making17
Idealizing analogy-making

abc ---> abd

mrrjjj ---> mrrjjk

Replace rightmost letter by successor

idealizing analogy making18
Idealizing analogy-making

abc ---> abd

mrrjjj ---> ?

idealizing analogy making19
Idealizing analogy-making

abc ---> abd

mrrjjj ---> ?

1 2 3

idealizing analogy making20
Idealizing analogy-making

abc ---> abd

mrrjjj ---> ?

1 2 3

1 2 4

idealizing analogy making21
Idealizing analogy-making

abc ---> abd

mrrjjj ---> mrrjjjj

1 2 3

Replace rightmost “letter” by successor

1 2 4

idealizing analogy making22
Idealizing analogy-making

abc ---> abd

xyz ---> ?

idealizing analogy making23
Idealizing analogy-making

abc ---> abd

xyz ---> xya

Replace rightmost letter by successor

idealizing analogy making24
Idealizing analogy-making

abc ---> abd

xyz ---> xya (not allowed)

Replace rightmost letter by successor

idealizing analogy making25
Idealizing analogy-making

abc ---> abd

xyz ---> ?

idealizing analogy making26
Idealizing analogy-making

abc ---> abd

xyz ---> ?

last letter in alphabet

idealizing analogy making27
Idealizing analogy-making

first letter in alphabet

abc ---> abd

xyz ---> ?

last letter in alphabet

idealizing analogy making28
Idealizing analogy-making

first letter in alphabet

abc ---> abd

xyz ---> ?

last letter in alphabet

idealizing analogy making29
Idealizing analogy-making

first letter in alphabet

abc ---> abd

xyz ---> wyz

last letter in alphabet

Replace “rightmost” letter by “successor”

idealizing analogy making30
Idealizing analogy-making

first letter in alphabet

abc ---> abd

xyz ---> wyz

last letter in alphabet

the copycat program hofstadter and mitchell1
The Copycat program(Hofstadter and Mitchell)
  • Inspired by collective behavior in complex systems (e.g., ant colonies)
the copycat program hofstadter and mitchell2
The Copycat program(Hofstadter and Mitchell)
  • Inspired by collective behavior in complex systems (e.g., ant colonies)
  • Understanding and perception of similarity is built up collectively by many independent simple “agents” working in parallel
the copycat program hofstadter and mitchell3
The Copycat program(Hofstadter and Mitchell)
  • Inspired by collective behavior in complex systems (e.g., ant colonies)
  • Understanding and perception of similarity is built up collectively by many independent simple “agents” working in parallel
  • Each agent has very limited perceptual and communication abilities
the copycat program hofstadter and mitchell4
The Copycat program(Hofstadter and Mitchell)
  • Inspired by collective behavior in complex systems (e.g., ant colonies)
  • Understanding and perception of similarity is built up collectively by many independent simple “agents” working in parallel
  • Each agent has very limited perceptual and communication abilities
  • Teams of agents explore different possibilities for structures, building on what previous teams have constructed.
the copycat program hofstadter and mitchell5
The Copycat program(Hofstadter and Mitchell)
  • Inspired by collective behavior in complex systems (e.g., ant colonies)
  • Understanding and perception of similarity is built up collectively by many independent simple “agents” working in parallel
  • Each agent has very limited perceptual and communication abilities
  • Teams of agents explore different possibilities for structures, building on what previous teams have constructed.
  • The resources (agent time) allocated to a possible structure depends on its promise, as assessed dynamically as exploration proceeds.
the copycat program hofstadter and mitchell6
The Copycat program(Hofstadter and Mitchell)
  • Inspired by collective behavior in complex systems (e.g., ant colonies)
  • Understanding and perception of similarity is built up collectively by many independent simple “agents” working in parallel
  • Each agent has very limited perceptual and communication abilities
  • Teams of agents explore different possibilities for structures, building on what previous teams have constructed.
  • The resources (agent time) allocated to a possible structure depends on its promise, as assessed dynamically as exploration proceeds.
  • The agents working together produce an “emergent” understanding of the analogy.
slide67

Architecture of Copycat

Concept network (Slipnet)

slide68

Architecture of Copycat

Concept network (Slipnet)

Workspace

a b c ---> a b d

i i j j k k --> ?

slide69

Architecture of Copycat

Concept network (Slipnet)

a b c ---> a b d

i i j j k k --> ?

Perceptual and structure-building agents (codelets)

Workspace

slide70

Architecture of Copycat

Concept network (Slipnet)

a b c ---> a b d

i i j j k k --> ?

Perceptual and structure-building agents (codelets)

Workspace

Temperature

slide72

Workspace

  • The Workspace starts out with letters in the analogy problem and their initial descriptions.
slide73

a b c --> a b d

m r r j j j --> ?

slide74

B

middle

letter

D

rightmost

letter

A

leftmost

letter

B

middle

letter

A

leftmost

letter

C

rightmost

letter

a b c --> a b d

M

leftmost

letter

R

letter

R

letter

J

letter

J

letter

J

leftmost

letter

m r r j j j --> ?

slide75

Workspace

  • The Workspace starts out with letters in the analogy problem and their initial descriptions.
  • Codelets gradually build up additional descriptions and structures.
slide76

a b c --> a b d

m r r j j j --> ?

slide77

a b c --> a b d

m r r j j j --> ?

slide78

Workspace

  • Codelets can be either “bottom-up” (noticers) or “top-down” (seekers).
slide79

a b c --> a b d

m r r j j j --> ?

slide80

successorship

a b c --> a b d

m r r j j j --> ?

slide81

successorship

a b c --> a b d

m r r j j j --> ?

slide82

Workspace

  • Codelets make probabilistic decisions:
slide83

Workspace

  • Codelets make probabilistic decisions:
    • What to look at next
slide84

Workspace

  • Codelets make probabilistic decisions:
    • What to look at next
    • Whether to build a structure there
slide85

Workspace

  • Codelets make probabilistic decisions:
    • What to look at next
    • Whether to build a structure there
    • How fast to build it
slide86

Workspace

  • Codelets make probabilistic decisions:
    • What to look at next
    • Whether to build a structure there
    • How fast to build it
    • Whether to destroy an existing structure there
slide87

a b c --> a b d

m r r j j j --> ?

slide88

a b c --> a b d

m r r j j j --> ?

rightmost --> rightmost??

letter --> letter??

slide89

a b c --> a b d

m r r j j j --> ?

rightmost --> rightmost?

letter --> letter?

slide90

a b c --> a b d

m r r j j j --> ?

leftmost --> leftmost??

letter --> letter??

rightmost --> rightmost?

letter --> letter?

slide91

a b c --> a b d

m r r j j j --> ?

leftmost --> leftmost??

letter --> letter??

rightmost --> rightmost

letter --> letter

slide92

a b c --> a b d

m r r j j j --> ?

leftmost --> leftmost??

letter --> letter??

rightmost --> rightmost

letter --> letter

slide93

a b c --> a b d

m r r j j j --> ?

leftmost --> leftmost??

letter --> letter??

rightmost --> rightmost

letter --> letter

rightmost --> rightmost??

letter --> group??

slide94

a b c --> a b d

m r r j j j --> ?

leftmost --> leftmost??

letter --> letter??

rightmost --> rightmost

letter --> letter

rightmost --> rightmost?

letter --> group?

slide95

a b c --> a b d

m r r j j j --> ?

leftmost --> leftmost?

letter --> letter?

rightmost --> rightmost

letter --> group

slide96

a b c --> a b d

m r r j j j --> ?

leftmost --> leftmost?

letter --> letter?

rightmost --> rightmost

letter --> group

leftmost --> rightmost??

letter --> letter??

slide97

a b c --> a b d

high

prob.

low

prob.

m r r j j j --> ?

leftmost --> leftmost?

letter --> letter?

rightmost --> rightmost

letter --> group

leftmost --> rightmost??

letter --> letter??

slide98
Probabilities are used to insure that no possibilities are ruled out in principle, but that not all possibilities have to be considered.
slide99
Probabilities are used to insure that no possibilities are ruled out in principle, but that not all possibilities have to be considered.
  • These decisions rely on information being obtained as the run takes place, e.g., pressure from current activation of concepts and neighboring structures.
slide100
Probabilities are used to insure that no possibilities are ruled out in principle, but that not all possibilities have to be considered.
  • These decisions rely on information being obtained as the run takes place, e.g., pressure from current activation of concepts and neighboring structures.
  • Therefore, the probabilities have to be updated continually.
slipnet1
Slipnet
  • Concepts are activated as instances are noticed in workspace.
slide104

a b c --> a b d

x y z --> ?

slide105

a b c --> a b d

x y z --> ?

slide106

successor

Part of Copycat’s Slipnet

slipnet2
Slipnet
  • Activation of concepts feeds back into “top-down” pressure to notice instances of those concepts in the workspace.
slide108

successor

a b c --> a b d

x y z --> ?

slide109

successor

a b c --> a b d

x y z --> ?

slide110

successor

a b c --> a b d

x y z --> ?

slipnet3
Slipnet
  • Activated concepts spread activation to neighboring concepts.
slide112

last

a b c --> a b d

last

x y z --> ?

slide113

first

last

a b c --> a b d

last

x y z --> ?

slide114

first

last

first

a b c --> a b d

last

x y z --> ?

slipnet4
Slipnet
  • Activation of link concepts determines current ease of slippages of that type (e.g., “opposite”).
slide116

rightmost

first

a b c --> a b d

leftmost

last

x y z --> ?

slide117

rightmost

first

a b c --> a b d

leftmost

last

x y z --> ?

slide118

rightmost

first

a b c --> a b d

leftmost

last

x y z --> ?

first --> last

opposite

last

first

slide119

leftmost

rightmost

rightmost

first

a b c --> a b d

leftmost

last

x y z --> ?

first --> last

opposite

last

first

slide120

leftmost

rightmost

rightmost

first

a b c --> a b d

leftmost

last

x y z --> ?

first --> last

opposite

last

first

slide121

leftmost

rightmost

rightmost

first

a b c --> a b d

leftmost

last

x y z --> ?

first --> last

rightmost --> leftmost

opposite

last

first

slide123
Measures how well organized the program’s “understanding” is as processing proceeds (a reflection of how good the current worldview is)

Temperature

slide124
Measures how well organized the program’s “understanding” is as processing proceeds (a reflection of how good the current worldview is)

Little organization —> high temperature

Temperature

slide125
Measures how well organized the program’s “understanding” is as processing proceeds (a reflection of how good the current worldview is)

Little organization —> high temperature

Lots of organization —> low temperature

Temperature

slide126

a b c --> a b d

m r r j j j --> ?

rightmost --> rightmost??

letter --> group??

leftmost --> leftmost?

letter --> letter?

High temperature

slide127

a b c --> a b d

m r r j j j --> ?

rightmost --> rightmost

letter --> group

leftmost --> leftmost?

letter --> letter?

Medium temperature

slide128

a b c --> a b d

m r r j j j --> ?

leftmost --> leftmost

letter --> group

rightmost --> rightmost

letter --> group

middle --> middle

letter --> group

Low temperature

slide129
Measures how well organized the program’s “understanding” is as processing proceeds (a reflection of how good the current worldview is)

Little organization —> high temperature

Lots of organization —> low temperature

Temperature

slide130
Measures how well organized the program’s “understanding” is as processing proceeds (a reflection of how good the current worldview is)

Little organization —> high temperature

Lots of organization —> low temperature

Temperature feeds back to codelets:

Temperature

slide131
Measures how well organized the program’s “understanding” is as processing proceeds (a reflection of how good the current worldview is)

Little organization —> high temperature

Lots of organization —> low temperature

Temperature feeds back to codelets:

High temperature —> low confidence in decisions —> decisions are made more randomly

Temperature

slide132
Measures how well organized the program’s “understanding” is as processing proceeds (a reflection of how good the current worldview is)

Little organization —> high temperature

Lots of organization —> low temperature

Temperature feeds back to codelets:

High temperature —> low confidence in decisions —> decisions are made more randomly

Low temperature —> high confidence in decisions —> decisions are made more deterministically

Temperature

slide133
Measures how well organized the program’s “understanding” is as processing proceeds (a reflection of how good the current worldview is)

Little organization —> high temperature

Lots of organization —> low temperature

Temperature feeds back to codelets:

High temperature —> low confidence in decisions —> decisions are made more randomly

Low temperature —> high confidence in decisions —> decisions are made more deterministically

Result: System gradually goes from random, parallel, bottom-up processing to deterministic, serial, top-down processing

Temperature

what s needed to apply these ideas in real world problems1
What’s needed to apply these ideas in “real world” problems?
  • Much expanded repertoire of concepts
what s needed to apply these ideas in real world problems2
What’s needed to apply these ideas in “real world” problems?
  • Much expanded repertoire of concepts
  • Ability to generate temporary concepts on the fly
what s needed to apply these ideas in real world problems3
What’s needed to apply these ideas in “real world” problems?
  • Much expanded repertoire of concepts
  • Ability to generate temporary concepts on the fly

(e.g., abc ---> abd, ace ---> ?)

what s needed to apply these ideas in real world problems4
What’s needed to apply these ideas in “real world” problems?
  • Much expanded repertoire of concepts
  • Ability to generate temporary concepts on the fly

(e.g., abc ---> abd, ace ---> ?)

  • Ability to learn new permanent concepts
what s needed to apply these ideas in real world problems5
What’s needed to apply these ideas in “real world” problems?
  • Much expanded repertoire of concepts
  • Ability to generate temporary concepts on the fly

(e.g., abc ---> abd, ace ---> ?)

  • Ability to learn new permanent concepts

(e.g., bbb ---> ddd, ppp ---> ?)

what s needed to apply these ideas in real world problems6
What’s needed to apply these ideas in “real world” problems?
  • Much expanded repertoire of concepts
  • Ability to generate temporary concepts on the fly

(e.g., abc ---> abd, ace ---> ?)

  • Ability to learn new permanent concepts

(e.g., bbb ---> ddd, ppp ---> ?)

  • Ability for “self-watching”
what s needed to apply these ideas in real world problems7
What’s needed to apply these ideas in “real world” problems?
  • Much expanded repertoire of concepts
  • Ability to generate temporary concepts on the fly

(e.g., abc ---> abd, ace ---> ?)

  • Ability to learn new permanent concepts

(e.g., bbb ---> ddd, ppp ---> ?)

  • Ability for “self-watching”

(e.g., abc ---> abd, xyz ---> ?)

what s needed to apply these ideas in real world problems8
What’s needed to apply these ideas in “real world” problems?
  • Much expanded repertoire of concepts
  • Ability to generate temporary concepts on the fly

(e.g., abc ---> abd, ace ---> ?)

  • Ability to learn new permanent concepts

(e.g., bbb ---> ddd, ppp ---> ?)

  • Ability for “self-watching”

(e.g., abc ---> abd, xyz ---> ?)

(cf. Marshall, Metacat, Ph.D. Dissertation, Indiana University, 1998)

applications of ideas from copycat1
Applications of Ideas from Copycat
  • Letter recognition (McGraw, 1995, Ph.D Dissertation, Indiana University)
applications of ideas from copycat2
Applications of Ideas from Copycat
  • Letter recognition (McGraw, 1995, Ph.D Dissertation, Indiana University)
  • Natural language processing (Gan, Palmer, and Lua, Computational Linguistics 22(4), 1996, pp. 531-553)
applications of ideas from copycat3
Applications of Ideas from Copycat
  • Letter recognition (McGraw, 1995, Ph.D Dissertation, Indiana University)
  • Natural language processing (Gan, Palmer, and Lua, Computational Linguistics 22(4), 1996, pp. 531-553)
  • Robot control (Lewis and Lugar, Proceedings of the 22nd Annual Conference of the Cognitive Science Society, Erlbaum, 2000.)
applications of ideas from copycat4
Applications of Ideas from Copycat
  • Letter recognition (McGraw, 1995, Ph.D Dissertation, Indiana University)
  • Natural language processing (Gan, Palmer, and Lua, Computational Linguistics 22(4), 1996, pp. 531-553)
  • Robot control (Lewis and Lugar, Proceedings of the 22nd Annual Conference of the Cognitive Science Society, Erlbaum, 2000.)
  • Image understanding (Mitchell et al.)