The sample dialogs below
are actual sequences produced by interactions with Babble. In
fact these are taken from Babble's reqression test. The text
in blue shows the input to Babble, the text in red is Babble's
response. The text in black discusses the significance of the
passage. |
|
Ian gave Emma the
big red ball.
She liked it.
Who liked the ball?
Emma
What color is the ball?
red |
|
Sentences describing simple events, (in this
case a give and a like event) are very easy for Babble to
understand. Notice Babble has to resolve both She and it to
answer the first question.
|
|
What is
the capital of Texas?
austin |
|
Here Babble's stored knowledge is queried.
Babble doesn't agonize over capitalization. |
|
Jumbo is a elephant.
What size is Jumbo?
big
Jumbo is small.
What size is Jumbo?
small |
|
Babble's ability to transfer properties of a category to the
members is part of Babble's if-then reasoning system. However,
Babble will give priority to properties assigned to an individual
over the category's properties. |
|
The car
Fred drives Paul likes.
Who likes the car?
Paul
Paul drove the car that has the roof that leaks.
What leaks?
roof
What has the roof?
car |
|
These examples that have somewhat harder
syntaxes. The first sentence is especially tricky to figure
out who's doing the driving and who's doing the liking. |
|
The car
that Jeff drove to the store that sold electronics crashed.
? r/destination
What crashed?
car
What sold electronics?
store |
|
Another example with many clauses, one
of which Babble is complaining about. It doesn't know how to
interpret the destination in the to clause. The point of this
example is that even though Babble couldn't figure out the to
clause, it could still get meaning from the rest of the sentence. |
|
Dan drives
the big white car.
Karen drives the little red car.
Who drives the red car?
Karen
Who drives the big car?
Dan
Who drives the little green car?
I don't know. |
|
This example demonstrates Babble's ability
to differentiate three different cars based on their properties. |
|
Robyn
is the mother of Tyler.
Who is the child of Robyn?
Tyler |
|
This is another example of Babble's
if-then reasoning, in this case using its knowledge of the relationship
between mother and child. If-then knowledge is also represented
by tridbits. |
|
The last
name of John is Smith.
What is the name of John?
I don't know.
What is the last name of John?
smith |
|
Names are a special type of property.
Babble differentiates "last name" from the unqualified
attribute of name. |
|
Who
are you?
Babble
You are hot.
Who is hot?
Babble |
|
Babble knows who it is and can process
information about itself. |
|
Who am
I?
Karen
I drink tea.
Who drinks tea?
Karen |
|
Babble also knows about the person it
is talking to, who is required to log in. |