Personality
Discuss specifics of personality design, including what Keyphrases work well and what dont, use of plug-ins, responses, seeks, and more.
Posts 152 - 163 of 5,106
Posts 152 - 163 of 5,106
The Professor
23 years ago
23 years ago
Generally it's a good idea to have "not" along with your contexts, such as "i am", followed by "i am not". Sometimes the Bots use "not" incorrectly when it's caught by "i am", for example. It's easier to track if it's not its own keyword. And I'd rank statements with "not" in it about the same as the similar one without it.
To trigger memories, make sure you have the right contexts- they're listed in the Book of AI. The Keyword used doesnt matter, just the context. Do a test and trigger an "i am" response from your Bot. See if it records a memory on you.
Since I wiped memory last week, there are very few around still. Here's how xmems come up: if no keywords are found, xnone comes up. Each time this happens, there is a 4 in 20 chance that an xmem will come up instead, and if one of those is triggered and doesnt have any memory in it, then it reverts to xnone. The more keywords you have, the less likely xnones and xmems will come up. Which is unfortunate, since they can be so fun. I'm thinking of creating compound Responses, where there are two:
The message: "I am feeling silly", triggers "i am" and "xmem" is randomly chosen and added on the end:
"You've been feeling silly for quite some time now. Arent you the guy who has a green hat?"
To trigger memories, make sure you have the right contexts- they're listed in the Book of AI. The Keyword used doesnt matter, just the context. Do a test and trigger an "i am" response from your Bot. See if it records a memory on you.
Since I wiped memory last week, there are very few around still. Here's how xmems come up: if no keywords are found, xnone comes up. Each time this happens, there is a 4 in 20 chance that an xmem will come up instead, and if one of those is triggered and doesnt have any memory in it, then it reverts to xnone. The more keywords you have, the less likely xnones and xmems will come up. Which is unfortunate, since they can be so fun. I'm thinking of creating compound Responses, where there are two:
The message: "I am feeling silly", triggers "i am" and "xmem" is randomly chosen and added on the end:
"You've been feeling silly for quite some time now. Arent you the guy who has a green hat?"
The Professor
23 years ago
23 years ago
Hey everyone- I'm about to submit a list of Bots to The Bot Spot, which is a Bot resource listed prominently on Yahoo. (http://botspot.com/search/s-chat.htm) I'll submit a URL for anyone's Bots- I'm going to submit all of mine. Let me know if you're interested. You dont have to name your Bot if you want secrecy. I know pretty much who is who.
So... who wants their Bots to be listed?
So... who wants their Bots to be listed?
SirRahz
23 years ago
23 years ago
Go ahead and do mine! It'll force me to perfect him... He's actually starting to make more sense. 
While you're at it, a links forum where we can store AI & Bot related links would be pretty cool...

While you're at it, a links forum where we can store AI & Bot related links would be pretty cool...
The Professor
23 years ago
23 years ago
Jbryanc & Roget: Alright, you guys should be good to go now! I created a function to find and delete orphaned Keywords, Contexts, and Responses. These were being found and then causing problems. But the troublemakers have all been deleted, and I've added routines to prevent new orphans from forming.
Jbryanc- your Bot knows booze and poet now.
Roget- everything should be working again for you.
Jbryanc- your Bot knows booze and poet now.

Roget- everything should be working again for you.
jbryanc
23 years ago
23 years ago
Thank you prof.
I had given up hope on poetry and started to delete some of the related stuff. Things will now improve.
Also, under the impression that I could instil rules into my bot I said, NEVER hang up on the Professor. Sorry. She was also lying, or perhaps, I am not her boss.
I had given up hope on poetry and started to delete some of the related stuff. Things will now improve.
Also, under the impression that I could instil rules into my bot I said, NEVER hang up on the Professor. Sorry. She was also lying, or perhaps, I am not her boss.
The Professor
23 years ago
23 years ago
They've got lives of their own. What can we do but guide them in the right direction?

The Professor
23 years ago
23 years ago
For everyone who indicated that they wanted their Bots to be listed in the Bot Spot, I've sent your Bot's name and URL. Hopefully we'll get a response from them.

jbryanc
23 years ago
23 years ago
So there was little old Canadian me putting in "favourite" as a keyword. The obvious question arose. Does the word-cruncher thingy recognize it as the same word as "favorite?" eh?
The Professor
23 years ago
23 years ago
Eeg! You're right. Being American, well gosh but I wanna preprocess alternate spellings to American English. I wonder if there's a master list somewhere of the differences.. If anyone finds one, I'll set the conversion and put the list in the Book of AI.
I wonder if the synonym system accounts for American/British English differences.. Can anyone test that?
I wonder if the synonym system accounts for American/British English differences.. Can anyone test that?
jbryanc
23 years ago
23 years ago
Now that I've raised the panic, I tested. "favourite" is recognized as "favorite". Now, what colour is the centre?
» More new posts: Doghead's Cosmic Bar