Visual Word Recognition - PowerPoint PPT Presentation

1 / 61
About This Presentation
Title:

Visual Word Recognition

Description:

Regularity vs. Consistency (Cortese & Simpson, 2000) Irregular words are usually irregular at two levels 1. grapheme-to-phoneme ... Michael Created Date: – PowerPoint PPT presentation

Number of Views:291
Avg rating:3.0/5.0
Slides: 62
Provided by: cort90
Category:

less

Transcript and Presenter's Notes

Title: Visual Word Recognition


1
Visual Word Recognition
2
Connectionist Models of Word Recognition
3
pint
The Dual Route Model
Letter Detectors
Lexicon (Mental Dictionary)
Sublexical System (Rules)
Pronunciation Code
4
Acquired Dyslexia Surface Dyslexia impaired
irregular-word reading pint, chaos
5
Acquired Dyslexia Surface Dyslexia impaired
irregular-word reading pint,
chaos Phonological Dyslexia impaired nonword
reading blark, flirp
6
The Dual Route Model Surface Dyslexia
pint
Letter Detectors
DAMAGE
Sublexical System (Rules)
Pronunciation Code
7
blask
The Dual Route Model Phonological Dyslexia
Letter Detectors
DAMAGE
Lexicon (Mental Dictionary)
Pronunciation Code
8
The PDP Model
Context
Meaning
Phonology (sound)
Spelling
pint
9
(No Transcript)
10
Properties of PDP models Representations are
distributed
pint /pInt/
11
Properties of PDP models Representations are
distributed
punt /punt/
pint /pInt/
12
Properties of PDP models Representations are
distributed
bank /bAnk/ side of a river
bank /bAnk/ money institution
13
Properties of PDP models Representations are
distributed
berry /beri/
bury /beri/
14
Properties of PDP models Representations are
distributed
bass /bAs/
bass /bas/
15
Properties of PDP models Processing is conducted
in parallel
16
Properties of PDP models The model learns
17
Properties of PDP models The model learns
tip /tip/
18
Properties of PDP models The model learns
19
Properties of PDP models The model learns
20
Properties of PDP models The model learns
21
Properties of PDP models The model learns
22
Properties of PDP models The model learns
23
Properties of PDP models Processing is
interactive
24
The Regularity Effect longer naming latencies
for irregular words than regular words Cortese
(1998) RTs in ms (error rates in
parentheses) pant 561 (.01) pint 632 (.28)
25
pint
The Dual Route Model
Letter Detectors
Lexicon (Mental Dictionary)
Sublexical System (Rules)
Pronunciation Code
26
pant
The Dual Route Model
Letter Detectors
Lexicon (Mental Dictionary)
Sublexical System (Rules)
Pronunciation Code
27
pint pant
lexical route sublexical route lexical route
sublexical route
i
I
a
output system
28
Stage 1
pint pant
lexical route sublexical route lexical route
sublexical route
i
I
a
output system
29
Stage 2
pint pant
lexical route sublexical route lexical route
sublexical route
i
I
a
output system
30
Stage 3
pint pant
lexical route sublexical route lexical route
sublexical route
i
I
u
output system
31
Stage 4
pint pant
lexical route sublexical route lexical route
sublexical route
i
I
a
output system
32
The Regularity Effect longer naming latencies
for irregular words than regular words Cortese
(1998) RTs in ms (error rates in
parentheses) and 3 of cycles for DRC
model people DRC model pant 561
(.01) 78.9 pint 632 (.28) 86.7
33
Regularity Effect in PDP Model hint mint weights
adjusted tint so that int /Int/ lint pint we
ights adjusted so that int /Int/ Cortese
(1998) - PDP error scores Regular (pant)
.070 Irregular (pint) .176
34
The PDP Model Surface Dyslexia
Context
DAMAGE
Phonology (sound)
Spelling
pint
35
pant
The Dual Route Model
Letter Detectors
Lexicon (Mental Dictionary)
Sublexical System (Rules)
Pronunciation Code
36
The Lexicon
Orthographic Lexicon
Semantic System
Phonological Lexicon
37
The PDP Model Phonological Dyslexia
Context
Meaning
DAMAGE
Phonology (sound)
Spelling
blask
38
Regularity vs. Consistency (Cortese Simpson,
2000) Irregular words are usually irregular at
two levels 1. grapheme-to-phoneme (e.g., pint) -
stick, lid, dish
39
Regularity vs. Consistency (Cortese Simpson,
2000) Irregular words are usually irregular at
two levels 1. grapheme-to-phoneme (e.g., pint) -
stick, lid, dish 2. word body (e.g., pint) -
mint, lint, tint
40
Regularity vs. Consistency (Cortese Simpson,
2000) Irregular words are usually irregular at
two levels 1. grapheme-to-phoneme (e.g., pint) -
stick, lid, dish 2. word body (e.g., pint) -
mint, lint, tint Consistency Effects are
Independent of GPC Regularity Regular Words -
moon vs. mood Glushko (1979)
41
Regularity vs. Consistency (Cortese Simpson,
2000) Irregular words are usually irregular at
two levels 1. grapheme-to-phoneme (e.g., pint) -
stick, lid, dish 2. word body (e.g., pint) -
mint, lint, tint Consistency Effects are
Independent of GPC Regularity Regular Words -
moon vs. mood Glushko (1979)
42
moon
The Dual Route Model
Letter Detectors
Lexicon (Mental Dictionary)
Sublexical System (Rules)
Pronunciation Code
43
mood
The Dual Route Model
Letter Detectors
Lexicon (Mental Dictionary)
Sublexical System (Rules)
Pronunciation Code
44
mood
The Dual Route Model
Letter Detectors

Sublexical System (Rules)
good stood wood hood mood food
Pronunciation Code
45
moon
The Dual Route Model
Letter Detectors

Sublexical System (Rules)
goon boon noon soon moon spoon
Pronunciation Code
46
pint
The Dual Route Model
Letter Detectors
Lexicon (Mental Dictionary)
Sublexical System (Rules)
Pronunciation Code
47
Regularity Consistency Effect in PDP
Model good wood weights adjusted hood so that
ood /ud/ stood mood weights adjusted so
that ood /ood/

48
Regular Words   batch
spook FRIENDS ENEMIES FRIENDS
ENEMIES patch watch kook
book catch cook latch
look match took hatch
shook rook hook
 
49
Regular Words   batch
spook FRIENDS ENEMIES FRIENDS
ENEMIES patch watch kook
book catch cook latch
look match took hatch
shook rook hook
Irregular Words   bind
pint   FRIENDS ENEMIES FRIENDS
ENEMIES kind wind
mint mind tint find
lint blind flint hind
print wind stint splint
 
50
batch
The Dual Route Model Regular Consistent
FgtE
Letter Detectors

Sublexical System (Rules)
match hatch batch latch catch
watch
Pronunciation Code
51
spook
The Dual Route Model Regular
Inconsistent EgtF
Letter Detectors

Sublexical System (Rules)
cook look took rook shook spook
Pronunciation Code
52
bind
The Dual Route Model Irregular
Consistent FgtE
Letter Detectors
mind blind hind kind find
bind
Sublexical System (Rules)
Pronunciation Code
-
53
pint
The Dual Route Model Irregular
Inconsistent EgtF
Letter Detectors
mint tint lint print pint
Sublexical System (Rules)
Pronunciation Code
-
54
Regular Words   batch
spook FRIENDS ENEMIES FRIENDS
ENEMIES patch watch kook
book catch cook latch
look match took hatch
shook rook hook
Irregular Words   bind
pint   FRIENDS ENEMIES FRIENDS
ENEMIES kind wind
mint mind tint find
lint blind flint hind
print wind stint
splint
 
55
pint
FgtE EgtF
spook
bind
Word Naming Time
What People Do
batch
56
Processing Time
The PDP Model
57
  • Summary
  • word-body consistency has a larger effect than
  • GPC regularity.

58
  • Summary
  • word-body consistency has a larger effect than
  • GPC regularity.
  • 2. The data are consistent the PDP model, and
    they are
  • inconsistent with the DRC model.
  • PDP Wins!

59
(1997)
60
pint
The Dual Route Model
Letter Detectors
Lexicon (Mental Dictionary)
Sublexical System (Rules)
Pronunciation Code
61
(2007)
Write a Comment
User Comments (0)
About PowerShow.com