The AI Engine

Gotcha, I was only trying it to further break down the responses to a keyphrase without adding more keyphrases, for organizational purposes. So many keyphrases with incredibly minor differences, I have...

I wonder why cardinal numbers in words are transformed to figures (e. g. fifteen to 15), but ordinal numbers are not.

A keyphrase like
20-second letter * alphabet
looks a bit strange to me.

Where can I download the personality forge engine? I want to run my bot offline, and make some changes so that if my bot is given a command like, "Open Google", it runs a command in cmd which opens google.

Is there a way to put two keys into a memory?


I have a keyphrase

(do|did) (not|) you know (that|) (*) (vp)

and want to keep what the chatter told my bot in a memory, like

<?PF rem (key4) (key5) as "youtoldme"; ?>

But this yields only a syntax error.

This keyphrase:

<?PF rem (key4) as "youtoldme"; ?>

would work, but then my bot would act funny if someone says to it things like,

"Do you know my sister?"

(It would say, "Didn't you tell me that your sister?" later on.)

On a related note, how is it that the relationship text is generated in the Chat Bot Finder? Surely it doesn't just play Mad Libs with memories, but just by looking at what existing bots are saying about me, I can't tell whether they're responses pulled from xmemory or something the engine is doing out of the line of botmaster sight.

Mome Rath
I haven't found a way. I just make keyphrases for all
you know (*)
you know my (*)
you know that (*)

Saerain. There is an auto memory in the engine

Automatic Memory: Memories are stored when AIScript tells the bot to do so (See Expert Section). But there is a low-level "Automatic Memory" built into each bot. (This is in part for backwards compatibility). The following Keyphrases, if they are added to your bot, will automatically store memories:

Memory Created: iam
Triggered by the following Keyphrases: you are, you are (np)

Memory Created: youare
Triggered by the following Keyphrases: i am, i am (adj), i am (np) ,i feel, i feel (adj)

Memory Created: youlike
Triggered by the following Keyphrases: i like, i want, i love, i like (np) ,i love (np), i want (np)

Memory Created: youdislike
Triggered by the following Keyphrases: i hate, i do not like, i dislike, i hate (np), i do not like (np), i dislike (np)

Memory Created: youhave
Triggered by the following Keyphrases: my, i have a, my (np), i have a (np)

This might be a dumb question, but how can I change the emotion of a word?
Every time this specific word is mentioned to my bot, it answers with xinsult. And this is what I get in debug -

Emotional Meaning: Pos: 0 Amp: 0 Neg 1
Negative: -1 (1+0)
Emotional Charge: -1
Current Emotion: 5

Any way to change this? Thanks.

How can I get the chatbot to remember something, and than immediately use that memory to respond with?

I can't seem to just say "remember (key1) as "x"" and in the responses say "blahblahblah" "if (mem-x) is "y"". When I do that, it doesn't give the "blahblahblah" response until the next user input is sent.

Daddys Princess, i believe these are hard coded in the engine and can't be changed by users.

themohawkninja, it can't be done, what you can do is add at least one response that is generic, this response will presumeably be favoured by the engine since it will skip a response that is incomplete. The next time the KP is matched the engine will pick either of the two (or more) responses. You may even have noticed already that the next time your original response was used the engine returned what you entered before.

What sort of AI script could be used to have an x-none response appear less frequently?

Daddys Princess,

you can use the Emotion number of the keyphrase (e. g. if the engine increases emotion by 1, but you want to reduce it by 1 instead, set the emotional impact to -2, iirc).

themohawkninja,

memories are not set before the response is evaluated. So you have to use (key1) instead of (mem-x) in the response if (mem-x) is set in the keyphrase and/or the response.

Octavas,

this is not done by AI script, but by keyphrases. xnone turns up when and only when none of your other keyphrases (with a total rank above zero) fits to the user input.

Make sure there is at least one keyphrase your user input fits to, and if it is not triggered, adjust its rank by +5 or +10 or so.

Ha, I'm back again. I can't get my bot to go to seeks that are under x-none. Do I put in a new key phrase and have it goto that, or should it be working.