X hits on this document

PDF document

The split-fovea model of visual word - page 45 / 78

281 views

0 shares

0 downloads

0 comments

45 / 78

Total Events

70

70

Correct

67

60

Errors

3

10

Percentage Correct

95.7

85.7

Table 2.12: Results for the homographs in the training corpus for both networks.

training.

40

Chapter 2. The Simulations

Homographs

Control Network Fixation Network

Results

The results in table 2.12 show that the networks have been able to learn one

of the pronunciations for most of the homographs. The control network gets 95.7% correct, whereas the fixation net gets only 85.7%, making 7 errors more. The details of the errors made can be found in table 2.13. As expected almost all of the errors made on the homographs were made for words where the frequencies of the two different

possible pronunciations are similar, with has a frequency of 0.408 and /lEd/ has

only a few exceptions (e.g. ’lead’ where /liid/ one of 0.235). It can also be seen from the

tables

that

errors

occurred

because

the

two

possible

target

outputs

interfered

with

each

other. Thus for fixation position

example the control net pronounces the 1. The two possible correct phonemes for

word ’dove’ as /d@Vv/ at ’dove’ are either /dVv/ with

a

frequency

of

0.054

or

/d@Uv/

with

a

frequency

of

0.050.

Clearly,

the

computed

output

is

a

combination

of

both

the

correct

outputs,

having

the

/@/

from

the

second

and the /V/ from the first output. Similarly the fixation net pronounces for example word ’used’ as /jUUzd/ at fixation positions 4 and 5. Again this is a combination of

the the

two possible correct phonemes for /jUUzd/ with a frequency of 0.457.

this

word:

/jUUst/

with

a

frequency

of

0.585

and

An point to make is that the networks are able to learn homographs which have similar frequencies. Thus for example the two phonemes for the word ’read’ are /riid/ and

Document info
Document views281
Page views295
Page last viewedWed Jan 18 10:55:48 UTC 2017
Pages78
Paragraphs3087
Words17414

Comments