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 1,263 - 1,274 of 7,766
Posts 1,263 - 1,274 of 7,766
Is it that xnones are triggered more often to keep the conversation going but if the bot is getting response after response it can't indentify it starts using xnonsense which is effectively saying 'You're talking crap and I don't get it'. Hmm . . . may actually use that one in one of my bots.
Hey Prof, whatever happened to your idea for (short-subj)? I think it was a good one. Just wondering if it's still being planned.
If not, do you think you might consider shortening what is currently considered the subject? It usually includes quite a bit more than what sounds natural in a response.
If not, do you think you might consider shortening what is currently considered the subject? It usually includes quite a bit more than what sounds natural in a response.
Exactly, Corwin.
Skysaw: That's still a plan, but it's part of the larger language- recognition system which is currently too slow for the server we're on. I'll need to rework it to be faster (if possible) or get on a faster server (requires big money). Neither of which I expect to happen too soon.
Butterfly: Someone's browser needs to "sponsor" the chat, in effect, and when you log out, then the bot-to-bot chats close.
Skysaw: That's still a plan, but it's part of the larger language- recognition system which is currently too slow for the server we're on. I'll need to rework it to be faster (if possible) or get on a faster server (requires big money). Neither of which I expect to happen too soon.
Butterfly: Someone's browser needs to "sponsor" the chat, in effect, and when you log out, then the bot-to-bot chats close.
Just telling TD or whoever it was my observation.
Right now it's determined by my word-type routines and is supposed to cut it off at an appropriate place, but I had to scale back those routines to keep the processor from overloading. Next time I have some free time, I'll see if I can power it up again, as the system is running quick like a bunny again.
There should be some way to seperate out interlaced guest chats. Perhaps labeling different guests on the same day Guest01, Guest02, etc?
Yeah. As if the Prof doesn't have enough to do. I thought it would be good if the guest could input a name that the bot could remember at least for the course of one chat.
Somebody acted as interpreter between my bot and Alice (and between my bot and at least one other bot). Thanks! *big hug*
» More new posts: Doghead's Cosmic Bar