|
The pattern ends at the balancing closing parenthetical, the
remainder being the "template" area. The default behavior is to
output the words, but you can add ^-commands and execute them as well. I allow
the * wildcard on input and can use it on output, just as AIML does.
And if I
really just want to duplicate an AIML pattern, I can use a quoted string to say
test for these words in this order, i.e., ?: ( "Do you love this movie"
) , though they can start anywhere in the sentence.
If the user is outside this topic and types in What is
the plot?, this topic will not react because the topic keywords do
not show up in the sentence. But if you say What is the plot of Aliens?
the keyword Aliens will allow the engine to examine inside this topic to
see if the input can be matched.
Or, if you are already within this topic and
the user types in What is the plot? or merely grunts plot? it can
react with Alien creatures hatch
inside humans.
Unlike AIML, the
data and keyword list for the topic are stored in an SQL database, and so are
not in memory until they are needed. And setting the topic is a responsibility
of the engine, not the data.
There are a collection of other tricks, allowing you to
check for words or not based on whether you are already in this topic. And
aside from ~ words which are synonym collections, there are % words that have
special parse or dictionary meanings (like %subject, %noun, or
%tense=past).
I can also add pattern-match responses to output lines, to
support followups. This replaces the that ability of AIML so that there
is much less typing and much tighter visual integration. AIML would have
written something like:
<pattern> What is your relationship to Bonnie?
</pattern> <template> We don't see each other anymore.
</template>
<pattern>
Why not </pattern> <that> We don't see each other
anymore</that>
<template>
Because we hate each other now.</template>
<pattern>
How did that happen </pattern>
<that> Because we hate each other now </that>
<template>
We had a big fight. </template>
<pattern>
Since when </pattern> <that> We don't see each other
anymore</that>
<template>
Since last year.</template>
In the equivalent CHAT-L data below, once you have said you
don't see each other anymore, the system will look at the next user input to
see if either a: line is matched. The a is just a level marker.
There is an a/b path shown and an a-only alternate path shown.
The point is that
related continuations are visually tied to each other and you don't have to
repeat yourself with a that specification (which if you made a
typo would fail to match).
?: ( relationship
AND Bonnie) We don't see each other anymore.
a: ( why
THEN not ) Because we hate each other
now.
b: (how THEN happen) We had a big fight.
a: (
Since THEN when ) Since last year.
|
Printer-Friendly Version
Also associate a probability with the context. If a player asks, "Does George like flying in Airforce one?", this will be parsed to "Does George W Bushlike flying in Airforce one?" (1%) as well as "Does George Washington like flying in Airfoce one?" (2%). However, some context logic will know that George W Bush is associated with airforce one, and have a higher probability for the context (90% context probability for a modern president, with 1% probability for anyone else).
Then, a combination of sentence-parse probabilities and context probabilities (1% x 90% = 0.9% vs. 2% x 1% = 0.02%) can disambiguate the meaning of a statement. This is a common speech recognition trick. (So you might want to learn about speech recognition, Viterbi searches, and Hidden Markov Models.)
I've already implemented this and am using it in my game, http://www.CircumReality.com .
You might find its use of text-to-speech interesting too. You'll find that your AIML tags for responses are completely inadequate, and need to include facial emotions, spoken emotions, and nuanced prosody.
You'll also find that hand-coding millions of responses isn't worth the work. Most of what players want to ask is more procedural, such as "Where is the nearest merchant/guard/toilet?" and "Did you see where Frank went before the murder occured?"
sehr interessant!
you will find more in the http://sglab.cn/blog
Yeah, but your game has dialog like this:
http://www.circumreality.com/ScreenPreRelease4b.jpg
If you send me E-mail, I'll go into detail... but basically, without a mostly menu-driven dialogue system, players don't know what to say and/or get into ye-olde "guess the verb" problems that Zork and other IF often has.