Language - Computational Explorations in Cognitive Neuroscience - page 365

Task Sentence

word, to speed things up a bit. The first word of the
active semantic role assignment sentence (schoolgirl)
is presented, and the network correctly answers that
schoolgirl is the agent of the sentence. Note that there is
no plus-phase and no training during this testing, so ev-
erything depends on the integration of the input words
— the small weight changes that we observed during
training cannot bias the correct answer.

Role assignment
Active semantic The schoolgirl stirred the Kool-Aid with a spoon.

Active syntactic The busdriver gave the rose to the teacher.

Passive semantic The jelly was spread by the busdriver with the knife.

Passive syntactic The teacher was kissed by the busdriver.

(control) The busdriver kissed the teacher.

Word ambiguity The busdriver threw the ball in the park.

The teacher threw the ball in the living room.

Concept instantiation The teacher kissed someone (male).

Role elaboration The schoolgirl ate crackers (with finger).

The schoolgirl ate (soup).

Online update The child ate soup with daintiness.

Continue to Step through to the final word in this
sentence (spoon).
You should observe that the network is able to iden-
tify correctly the roles of all of the words presented.
Because in this sentence the roles of the words are con-
strained by their semantics, this success demonstrates
that the network is sensitive to these semantic con-
straints and can use them in parsing.

(control) The pitcher ate soup with daintiness.

, !

Conflict The adult drank iced tea in the kitchen (living room).

Table 10.15: Test cases for exploring the behavior of the
model, as described in the text.

we will just load in some pretrained weights from a net-
work that was trained for 225 epochs of 100 sentences
per epoch.

Now Step through the next sentence.
This sentence has two animate nouns (busdriver and
teacher), so the network must use the syntactic word or-
der cues to infer that the busdriver is the agent, while
using the “gave to” syntactic construction to recognize
that the teacher is the recipient. Observe that at the final
word in the sentence, the network has correctly identi-
fied all the words.
In the next sentence, the passive construction is used,
but this should be obvious from the semantic cue that
jelly cannot be an agent.

Do LoadNet on the overall sg_ctrl control panel.
The test sentences are designed to illustrate different
aspects of the sentence comprehension task, as noted in
the table. First, a set of role assignment tasks provide ei-
ther semantic or purely syntactic cues to assign the main
roles in the sentence. The semantic cues depend on the
fact that only animate nouns can be agents, whereas
inanimate nouns can only be patients. Although ani-
macy is not explicitly provided in the input, the train-
ing environment enforces this constraint. Thus, when
a sentence starts off with “The jelly...,” we can tell that
because jelly is inanimate, it must be the patient of a
passive sentence, and not the agent of an active one.
However, if the sentence begins with “The busdriver...,”
we do not know if the busdriver is an agent or a patient,
and we thus have to wait to see if the syntactic cue of
the word was appears next.

, !

Step through and observe that the network correctly
parses this sentence.
In the final role assignment case, the sentence is pas-
sive and there are only syntactic constraints available to
identify whether the teacher is the agent or the patient.
This is the most difficult construction that the network
faces, and it does not appear to get it right — it gets
confused when busdriver appears at the end of the sen-
tence, and concludes that the busdriver is both the agent
and patient of the sentence.

, !

To start testing, iconify the training text log and
process control panel, and then open a testing log
( Trial_1_TextLog ) by doing View , TEST_LOG . Then,
open a testing process control panel ( Epoch_1 )using
View , TEST_PROCESS_CTRL .
The testing text log has the same columns as the train-
ing one (figure 10.28), so the results of testing can be
interpreted the same way.

Step through this sentence.
Further testing has shown that the network sometimes
gets this sentence right, but often makes errors. This can
apparently be attributed to the lower frequency of pas-
sive sentences, as you can see from the next sentence,
which is a “control condition” of the higher frequency
active form of the previous sentence, with which the
network has no difficulties.

, !

Do ReInit and Step in the process control panel.
Each step here covers all questions for each input

, !