KCP Maastricht

1
KCP Maastricht

• Expertise
• Jaap Murre
• jaap_at_murre.com

2
Example of expertise

• World champion blind simultaneous checkers Ton
  Sijbrands
• On 5 oktober 2007 won 21 games and played 4 at a
  draw, blind! (it took more than 28 hours)
• In 1982 only 10 games (9 won, 1 draw, 620 hours)
• How can someone reach this level of expertise?
  What is expertise?

3
Blind simultaneous checkers Ton Sijbrands
4
Five opponents (19 May 2007)
5
Extreme expertise
6
Beginners are often the hardest to keep track of!
7
Learning skills

• Three stages
• Cognitive stage
• declarative encoding (not procedural)
• Associative stage
• detect and eliminate errors
• both procedural and declarative knowledge
  (production rules, procedularization)
• Autonomous stage
• automization (unconscious of details)

8
expertise

• What happens when you gain expertise?
• You become better and faster
• (power law of practice)

9
Power Law of Practice

• power function (also exponential)
• P bt-a
• P achievement
• t time
• a learning constant
• b scaling constant
• A power function gives a straight line on a
  log-log plot

10
Pirolli Anderson 1985
Recognition of sentences studied earlier
Anderson 2000
11
Crossman (1959)
Cigar making
Anderson 2000
12
Kolers (1976)
Anderson 2000
13
Kolers (1976)
Anderson 2000
14
expertise

• tactical and strategic learning

15
tactical learning - tactics

• Repeat the same method to reach a goal
• The succession of actions is called tactical
  learning
• Logan (1988) you remember the solutions to
  specific problem instances

16
Logan (1988)

• D 3

(E, F, ) G
17
Logan (1988)

• F 5
• L 5
• P 5
• S 3
• C 1
• T 1
• J 2

• K
• Q
• U
• V
• D
• U
• L

18
1e
12e
19
strategic learning

• In what way do you organize the problem?
• programming (computer program)
• beginner depth first
• expert breadth first

20
computer program
21
computer program
Beginner
depth-first
22
computer program
Expert
breadth-first
23
strategic learning

• Learning how to solve a range of problems
• But how is domain specific

24
Adriaan de Groot 1965
Anderson 2000
25
Chess experts

• They see chunks or patterns and not isolated
  chess pieces
• These patterns are stored in long-term memory
• In memory is also store how to act with a
  specific pattern

26
Transfer of training?

• Latin school Difficult subjects will train the
  mind
• Does expertise transfer to another field or
  domain?

27
Problems with transfer!!

• tactical learning
• specific knowledge is stored in memory
• strategic learning
• organization of problem is domain specific

28
Hoe word je een expert?
29
Mere practice? -gtEricsson

• Ericsson expertise almost completely explained
  by practice
• Time on task

30
Excellent presentation by Ericsson at
the Symposium 100 Psychologie at the UvA. See
video and slides at http//www.100jaarpsychologi
e.uva.nl/
31
Slide derived from Ericssons Lecture at UvA 9
November 2007
32
Slide derived from Ericssons Lecture at UvA 9
November 2007
33
Slide derived from Ericssons Lecture at UvA 9
November 2007
34
Test yourself 1 per second
Slide derived from Ericssons Lecture at UvA 9
November 2007
35
Slide derived from Ericssons Lecture at UvA 9
November 2007
36
Slide derived from Ericssons Lecture at UvA 9
November 2007
37
Slide derived from Ericssons Lecture at UvA 9
November 2007
38
Slide derived from Ericssons Lecture at UvA 9
November 2007
39
Slide derived from Ericssons Lecture at UvA 9
November 2007
40
Derived from Ericssons Lecture at UvA 9
November 2007
41
Slide derived from Ericssons Lecture at UvA 9
November 2007
42
Slide derived from Ericssons Lecture at UvA 9
November 2007
43
Slide derived from Ericssons Lecture at UvA 9
November 2007
44
Slide derived from Ericssons Lecture at UvA 9
November 2007
45
Oefening??
Ericsson, Academie Berlijn 7000 vs. 5000 uur
oefening?
Paganini.
46
Slide derived from Ericssons Lecture at UvA 9
November 2007
47
Slide derived from Ericssons Lecture at UvA 9
November 2007
48
Slide derived from Ericssons Lecture at UvA 9
November 2007
49
Slide derived from Ericssons Lecture at UvA 9
November 2007
50
Slide derived from Ericssons Lecture at UvA 9
November 2007
51
Slide derived from Ericssons Lecture at UvA 9
November 2007
52
Slide derived from Ericssons Lecture at UvA 9
November 2007
53
Slide derived from Ericssons Lecture at UvA 9
November 2007
54
Slide derived from Ericssons Lecture at UvA 9
November 2007
55
Slide derived from Ericssons Lecture at UvA 9
November 2007
56
Slide derived from Ericssons Lecture at UvA 9
November 2007
57
Slide derived from Ericssons Lecture at UvA 9
November 2007
58
Slide derived from Ericssons Lecture at UvA 9
November 2007
59
Slide derived from Ericssons Lecture at UvA 9
November 2007
60
Slide derived from Ericssons Lecture at UvA 9
November 2007
61
Slide derived from Ericssons Lecture at UvA 9
November 2007
62
Brain plasticity and expertise

• There is now a lot of evidence that the brains
  structure changes due to practice
• How this is cause by gains in expertise is not
  investigated in much detail
• It will depend on the type of expertise (compare
  memory wonder, chess player, martial arts
  specialist, etc.)

63
Example plasticity in somatosensory cortex
64
Sensory Homunculus
65
Sensory maps
66
Changes in topological maps as a result of
altered experience
67
Slide derived from Ericssons Lecture at UvA 9
November 2007
68
Slide derived from Ericssons Lecture at UvA 9
November 2007
69
Slide derived from Ericssons Lecture at UvA 9
November 2007
70
Slide derived from Ericssons Lecture at UvA 9
November 2007
71
Slide derived from Ericssons Lecture at UvA 9
November 2007
72
Slide derived from Ericssons Lecture at UvA 9
November 2007
73
talent?

• It is clear that expertise only emerges after
  prolonged training (10 year rule)
• Ericsson et al. Almost all expertise through
  practice, not talent
• But deliberate practice!

74

• Intelligence

75
Fascination with IQ

• Coxs 301 Geniuses
• Estimation of intelligence
• real-life intelligence

76
Example Francis Galton

• 18 months knows alphabet
• 2.5 years reads a short book
• 3 years writes his name
• 4 years reads some French, can add and
  multiply, can tell time
• 7 years - reads Shakespeare for fun, can store a
  page in memory by reading it twice

77
IQ follows a normal distribution with standard
deviation of 15
78
Some estimates by Cox

• Napoleon 145
• Voltaire 190
• Goethe 210
• Mozart 165
• Darwin 165
• Michelangelo 180
• Isaac Newton 190
• Abraham Lincoln 150

Goethe
79
Coxs estimated IQ per profession

• Philosopher 180
• Scientist 175
• Fiction writers 165
• Politicians 165
• Artists 160

80
IQ - Intelligence?

• What is intelligence, what is IQ?
• Am I intelligent?
• Mensa accepts people with an IQ gt 131 (i.e., top
  2)

81
Mensa Test
82
(No Transcript)
83
(No Transcript)
84
High IQ Societies

• Top 2 (98th percentile IQ 130 sd15, IQ 132
  sd16) Mensa International, High Potentials
  Society, Mysterium Society, Altacapacidadhispana,
  SocratIQ Society, Encefálica
• Top 1 (99th percentile IQ 135 sd15, IQ 137
  sd16) Intertel, Top One Percent Society,
  Elateneo/s Society, Superdotados-Intelectuales,
  The Mind Society, Sinapsa Society
• Top 0.5 (99.5th percentile IQ 139 sd15, IQ 141
  sd16) Colloquy, Poetic Genius Society
• Top 0.2 (99.8th percentile IQ 143 sd15, IQ 146
  sd16) ePiq Society, Neurocubo

85
Very high IQ Societies

• Top 0.1 (99.9th percentile IQ 146 sd15, IQ 149
  sd16) International Society for Philosophical
  Enquiry, Triple Nine Society, IQuadrivium Society
  
• Top 0.09 (99.91th percentile IQ 147 sd15, IQ
  150 sd16) Glia Society, One-in-a-Thousand
  Society, Milenija
• Top 0.0001 (99.9999th percentile IQ 172 sd15,
  IQ 176 sd16) Mega Society, Mega International,
  Pi Society, Omega Society, StrictIQ Society

86
Extremely high IQ Societies

• Top 0.00003 (99.99997th percentile IQ 175 sd15,
  IQ 180 sd16) OLYMPIQ Society, PARS Society
• Top 0.0000001 (99.9999999th percentile IQ 190
  sd 15, IQ196 sd16) Giga Society

87
Giga the most exclusive IQ society

• Giga society was founded in 1996 by Paul
  Cooijmans, who has served as its Administrator
  since.
• Giga Society is open  to anyone scoring at  or 
  above the  99.9999999th estimated unselected
  population percentile on any of the acceptable
  tests.
• This means one in a  billion  individuals can
  qualify. The corresponding IQ, expressed  on  a
  scale with a set standard deviation of 16, is
  196, respectively sd 15, an IQ 190.
• There are currently seven Giga members.

88
Mega Society Test only 0.0001

• 1. STRIP MÖBIUS BOTTLE ?2. THOUGHT
  ACTION OBSESSIVE ?3. LACKING MONEY
  PENURIOUS DOTING ON ONE'S WIFE ?4. MICE
  MEN CABBAGES ?5. TIRE RETREAD
  PARCHMENT ?6. ALL IS ONE MONISM ALL IS
  SELF ?7. SWORD DAMOCLES BED
• 48.
• Marilyn vos Savant scored 46 correct IQ 186
  (wereldrecord)
• See http//www.eskimo.com/miyaguch/titan.html
  for all 48 questions (must have 43 right for
  Mega! But may take one month.)

89
Horn, Anderson pg. 456

• Although the word intelligence (as a unitary
  concept) continues to be useful in everyday life,
  this does not represent a good scientific concept

90
Intelligence testing

• Raven
• Tests general intelligence?
• Clearly influenced by general processes like
  working memory

91
Raven Progressive Matrices
92
Raven (difficult example)
93
Relation between subtests usually positive
Uit Jelte Wicherts lezing op 9 november 2007, UvA
94
A correlation is always found
Uit Jelte Wicherts lezing op 9 november 2007, UvA
95
Examples of subtests
Uit Jelte Wicherts lezing op 9 november 2007, UvA
96
All tests correlate positively
Uit Jelte Wicherts lezing op 9 november 2007, UvA
97
Positive Manifold
Uit Jelte Wicherts lezing op 9 november 2007, UvA
98
Uit Jelte Wicherts lezing op 9 november 2007, UvA
99
Spearman (1904)
Uit Jelte Wicherts lezing op 9 november 2007, UvA
100
Decomposition of intelligence
Uit Jelte Wicherts lezing op 9 november 2007, UvA
101
Factor analysis
Uit Jelte Wicherts lezing op 9 november 2007, UvA
102
Uit Jelte Wicherts lezing op 9 november 2007, UvA
103
Uit Jelte Wicherts lezing op 9 november 2007, UvA
104
Also see presentation by Jelte Wicherts at the
Symposium 100 Jaar Psychologie at the UvA
http//www.100jaarpsychologie.uva.nl/
105
g (general intelligence)
Uit Jelte Wicherts lezing op 9 november 2007, UvA
106
Usefulness of g to predict
Uit Jelte Wicherts lezing op 9 november 2007, UvA
107
g

• speed of processing? (energy)
• working memory? (space)

108
fluid - crystallized

• Cattell
• Fluid skill in reasoning and information
  processing (in new domains)
• Crystallized acquired knowledge
• Vocabulary knowledge is a good proxy for IQ

109
Horn, Anderson pg. 456

• Although the word intelligence (as a unitary
  concept) continues to be useful in everyday life,
  this does not represent a good scientific concept

110
IQ is not intelligence

• IQ and intelligentie are not the same thing!
• IQ practical
• e.g., school achievement
• Intelligence theoretical
• understanding cognitive functioning

111
Uit Jelte Wicherts lezing op 9 november 2007, UvA
112
Uit Jelte Wicherts lezing op 9 november 2007, UvA
113
Mental score as a function of age
114
Uit Jelte Wicherts lezing op 9 november 2007, UvA
115
Uit Jelte Wicherts lezing op 9 november 2007, UvA
116
Uit Jelte Wicherts lezing op 9 november 2007, UvA
117
Uit Jelte Wicherts lezing op 9 november 2007, UvA
118
Uit Jelte Wicherts lezing op 9 november 2007, UvA
119
Intelligence conclusion

• Intelligence is an important socio-cultural
  concept
• Rarely neutral!
• Measured as IQ, but IQ is not synonymous with
  intelligence
• Scientifically it is not very clear what
  intelligence is