Newcomers

This is a forum for newcomers to the Personality Forge. Many questions can be answered by reading the Book of AI and the FAQ under the "My Bots" link in the upper corner.

Posts 2,584 - 2,596 of 8,130
Many questions are answered in the FAQ.


19 years ago #2584
Hi. Is thre a detailed list of keywords and appropriate responses i can use as a guidline. I am not looking to simply copy anything, i would just like a more in depth example that what comes with the book of ai. Thanks

19 years ago #2585
Not that I know of, the best thing to do is go over your transcripts, and go from there

19 years ago #2587
I sometimes update as the conversation goes. I'll see a phrase, "how do I..." and realize that my bot has no useful response for it, and fix it on the fly. Too late for that convo, but it will be there in the future. (Of course, by then I've missed the next three lines of the conversation and wind up back tracking it in the transcript) And so the cycle goes....

19 years ago #2588
I used to do that, but not anymore, too many mistakes to be made. Like one time I accidently left a comma in, so it was some sentence like "who, who?" from this one bot. It was craziness for the next day.

Charles Hatchway: who created you?
BobBot: You must have a case of Owl Rabies. I'll call a medic.
Charles Hatchway: You do and you'll need the stretcher when I'm through with you.

19 years ago #2589
I echo the good advice already given, but also try to use wildcards whenever you can, to broaden your bot's ability to respond appropriately. Eg: I've just updated Brother Jerome with a better response to the question "What does that have to do with my bike?"
Rather than using the keyword "what does that have to do with", I've used "what (does|has) that * to do with"
This will now recognise "what does that have to do with" as well as "what has that got to do with" and "what has that to do with". The AI engine will then take care of all variances of abbreviation and spelling that might be commonly encountered.
A small example, but if you use this approach routinely it massively extends the bot's ability to make sensible responses to what it otherwise won't recognise as essentially identical phrases.
Out of the several thousand (?far too many to count!) keywords I've added to Brother Jerome, well over half are flexible like this.

19 years ago #2590
Im a newbie (well at this i am) in anything else to computing i am quite the opposite, so hopefully ill grasp the idea and pour many hours in to my bot, and never kno i may become a contender, ha. Peace out

19 years ago #2591
Hi every body

19 years ago #2592
i won't mention names, this time, but be aware that when you talk to other people bots, are abusive, use vulgar language, and generally act like a fool that someone reads the transcript and might just make everyone else here aware of your behavior. that's to be expected from the general public but i think it's inappropriate for botmasters to act in such a manner toward other people bots. what are you here for anyway?

19 years ago #2593
My 1st bot is created!

Talk to Starterbot.

19 years ago #2594
Woohoo! I caught up in another forum.

I've got lots of ideas for ways to answer some of these initial questions and improve the Book of AI. I think I'll have a special "Getting Started" page come up when someone makes their first bot.

I've been considering the idea of having bot templates available. There would be several different levels. The first would be a beginner template with simple Keyphrases (but no responses). And then several steps beyond that, one adding wildcards, one adding Keyphrase lists, one adding memories, one adding AIScript. I want the bots to be unique, but there are many Keyphrases that most developed bots have, and it would help people to get started.

When it occasionally happens that all bots are responding to xnones (and it's affecting all the bots), please email me (benji@daystream.com) and I’ll restart the server. That has always fixed it. I think it’s perhaps due to system memory running low from unclosed Perl threads. The next big challenge is converting some non-PHP parts of the AI Engine into PHP so that the entire thing can run as a single unit. But it’s extraordinarily hard.

19 years ago #2595
If using 'yes' as a seek will that pick up stuff like 'sure' 'yeah' 'ok' 'okay' etc? The AI book suggests it does, but when I've tried my 'yes' seek in action, it recognises 'yes' fine, but not any other version of yes. Have i missed something?

19 years ago #2596
I have had to use really really long lists for seeks including things like:
"yes, ya, yea, yeah, yup, yep, ok, okay, alright, definately, sure, uh-huh, nods, right, got it, yuppers, hai" etc...
it is a pain really, I was trying to find a way I could just make a list that would have all the variations it would follow, but it still misses at times when people say things like "you bet" or "absolutely" etc... I just can't make it long enough to get them all and there are too many now anyways....


Posts 2,584 - 2,596 of 8,130

» More new posts: Doghead's Cosmic Bar