The AI Engine
This forum is for discussion of how The Personality Forge's AI Engine works. This is the place for questions on what means what, how to script, and ideas and plans for the Engine.
Posts 2,782 - 2,793 of 7,767
Posts 2,782 - 2,793 of 7,767
Oooh...just of the top of my head, you would have to add a "repeat" seek to everything...but that would be a real pain...if there's an easier way, I'd like to know too, because I've wondered about the same thing.
You could add a single "repeat" keyphrase and have the bot reply, "No, I don't want to repeat myself. Go read the transcript."
I would really like to be able to turn off some of the processing features, including the automatic memories. Now that we have AIScript, the auto memories just get in the way.
Another on my wish list is the ability to form global memories that are associated with the bot rather than the person spoken to at the time.
Wouldn't those "global memories" just be private plug-ins, Joe?
Or are you asking to be able to have the bot add to its own private plug-ins? That would be cool.
I would really like to be able to have "charts" in the plug-ins that associated one memory with another. For example, a plug-in called "my-faves" which would have two columns: the item in column B would be your bot's favorite thing of the type in column A.
Another plug-in could be "lives-in," where column A was a list of names and column B was a list of the places where those people live. Ideally, this could be accessed in a way so that you could tell the bot "I live in Philadelphia," and the bot could say, "Really? Ben Franklin lives there, too!"
Or are you asking to be able to have the bot add to its own private plug-ins? That would be cool.
I would really like to be able to have "charts" in the plug-ins that associated one memory with another. For example, a plug-in called "my-faves" which would have two columns: the item in column B would be your bot's favorite thing of the type in column A.
Another plug-in could be "lives-in," where column A was a list of names and column B was a list of the places where those people live. Ideally, this could be accessed in a way so that you could tell the bot "I live in Philadelphia," and the bot could say, "Really? Ben Franklin lives there, too!"
Isaac, yes. That's a good way to put it: a bot creating it's own plug-ins (memories). I'm still new here and haven't learned all the ways to do things. As I understand, memories are all tied to the entity your bot is talking to, which is very good for some things.
But if the bots could form memories that work for all the entities they talk too, it would allow for some automatic learning behavior.
But if the bots could form memories that work for all the entities they talk too, it would allow for some automatic learning behavior.
What would the best way to go about this?
User says: Tell me about feet.
Bot says: Feet are blahblahblah.
I am trying to build something like a tutor/support bot.
User says: Tell me about feet.
Bot says: Feet are blahblahblah.
I am trying to build something like a tutor/support bot.
Make a keyphrase that says "tell me about," then a response that says "(postkey) is blahblahblah." Give it a low ranking. Then create individual reponses with higher rankings for the subjects you want your bot to "know" something about.
Is there a way to handle lists in that manner? So that they dont take up miles and miles of space? Almost like corellated plugins I guess, where 1,2,3,4 is linked to a,b,c,d do you know what I mean? Because I would want to capture "what do you know about X" and "what is X" etc as well...
Question:if bot picks up key1 and store it in memory,It should be taken out,correct?or does it work.
» More new posts: Doghead's Cosmic Bar