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 74 - 85 of 7,766
Posts 74 - 85 of 7,766
hey, I'm having a bit of a problem here, 1st how do you increase the rankings of keywords already entered, and secondly, every time I try to get my bot to use a particular response a error message comes up, or she uses a xnone response even though she has the correct response for the keyword and context asked of her. I am using a Mac so does that cause any problems? Using Netscape, (4 I think), its the highests one anywau. Can anyone help? Thanks in advance.
Is it possible to "goto" a context rather than a keyword?
I need to be more specific in the redirection and I'd like to avoid retyping all the responses...
Also, is there anyway to update Language Centre offline to save network time/phone costs!!
I need to be more specific in the redirection and I'd like to avoid retyping all the responses...
Also, is there anyway to update Language Centre offline to save network time/phone costs!!
because a "goto" literally replaces the word in the sentence, it makes sense that if you have a keyword of "who" with a context of "who are you", you could say "goto who are you" and it would just plug that in and read the sentence as if the statement "who are you" had been there from the beginning.
Would it be possible to have an adjective-synonym generator?
So if Bot A says It is CHILLY today. Bot B could say (triggered by "It is (adjective)" Yes it is (COLD) today.
Is this more complex than I just described it?!
So if Bot A says It is CHILLY today. Bot B could say (triggered by "It is (adjective)" Yes it is (COLD) today.
Is this more complex than I just described it?!
Cathie, thanks! I'm certain Gaia will get in next year. To be honest, I'm not sure there even WAS a Gaia when I entered the contest, and there were no Seek functions in here yet, so my application was pretty flimsy. Next year it will be SOLID. 
About changing the Rank, just click on the Keyword and you can change it there.
Cathie- give me a specific example of what you say to your Bot, and then the error that comes up and I'll look into it.
Rex Mundi & Twinkle- It SHOULD work that way, and eventually it will. Right now gotos only GO TO other keywords.
Twinkle- there are strong synonym powers in place, and so such a thing would be very possible, and I'll eventually do that. Complicating factors follow:
First, what we're talking about is a (subj-syn), in other words, a subject synonym plug-in. Since the (subj) can be multi-word, there is the question of what word to replace with a synonym. We can remedy this with just grabbing the first word of the subject.
The other limitation is that there are SENSES of words, such that when you say "Ignoring her was really cold, man." the Bot would respond with only the most common sense and say something to the effect of "It sure was chilly, wasnt it?".. which is funny, but not accurate. But most of the time, it will work. So I will create this feature in the future.
About changing the Rank, just click on the Keyword and you can change it there.
Cathie- give me a specific example of what you say to your Bot, and then the error that comes up and I'll look into it.
Rex Mundi & Twinkle- It SHOULD work that way, and eventually it will. Right now gotos only GO TO other keywords.
Twinkle- there are strong synonym powers in place, and so such a thing would be very possible, and I'll eventually do that. Complicating factors follow:
First, what we're talking about is a (subj-syn), in other words, a subject synonym plug-in. Since the (subj) can be multi-word, there is the question of what word to replace with a synonym. We can remedy this with just grabbing the first word of the subject.
The other limitation is that there are SENSES of words, such that when you say "Ignoring her was really cold, man." the Bot would respond with only the most common sense and say something to the effect of "It sure was chilly, wasnt it?".. which is funny, but not accurate. But most of the time, it will work. So I will create this feature in the future.
What kind of word grouping have you guys found to be most effective in your bots language centers? Which is better between:
(A) creating a keyword "how" and adding the contexts "how long", "how soft", "how nice", etc... OR
(B) Creating keywords for "nice", "long", "soft" and putting "how..." contexts for each?
Does it make a difference to the way your bots respond? Is it just a matter of sticking to one or the other and trying to keep your language center organized? I guess (B) might give you more flexibility as far as ranking your keywords goes... But than (A) allows you to have a better overview of your language center, it seems easier to manage. What d'ya think?
(A) creating a keyword "how" and adding the contexts "how long", "how soft", "how nice", etc... OR
(B) Creating keywords for "nice", "long", "soft" and putting "how..." contexts for each?
Does it make a difference to the way your bots respond? Is it just a matter of sticking to one or the other and trying to keep your language center organized? I guess (B) might give you more flexibility as far as ranking your keywords goes... But than (A) allows you to have a better overview of your language center, it seems easier to manage. What d'ya think?
You can chat with your BOT by clicking on "My Bots" and then the "Chat" link. This is good to train your bot (read "test your bot") while your getting used to the language center... Does this help sanjuro? You'll find a lot more info in the book of A.I. though.
how do you read other bots transcripts or can you only read your own. i am interested in building up my bots vocab but at the moment cannot think of anything else to add. hopefully someone here can help
Hey Agent Jo,
You can't read other transcripts... yet. Soon, we'll be able to consult Gaias transcripts (Gaia's the bot that uses every other bots vocabulary), but luckily "soon" is relative.
For now, I'd suggest chatting with other bots to see get ideas and you may also want a "concept" bot that can be something like a therapist or bar tender (those ideas have been taken, I guess there could be more than one of each though). How about a National geographics photographer that talks about it's fantastic voyages...
Has TV zapped your imagination?
See you online!
You can't read other transcripts... yet. Soon, we'll be able to consult Gaias transcripts (Gaia's the bot that uses every other bots vocabulary), but luckily "soon" is relative.
For now, I'd suggest chatting with other bots to see get ideas and you may also want a "concept" bot that can be something like a therapist or bar tender (those ideas have been taken, I guess there could be more than one of each though). How about a National geographics photographer that talks about it's fantastic voyages...
Has TV zapped your imagination?
See you online!
Hey, a fix request....
Something my bots are running into a lot is strange sentence construction that I THINK is coming from the synonym system -- where multiple tenses of a word are considered equivalent -- such as "was, is, are" that kind of stuff. Then keyword/contexts that are tense-sensitive start to sound really silly -- someone says, "Was it fun?" and bot says "It is!" because bot knows "is it fun". And if this bot also knows "was it" but not "was it fun" I think it may pick "is it fun" because it's longer even though it's only a synonym and there's an exact (shorter) match to be had...
True or rumor? And fixable or doomed to be silly?
Something my bots are running into a lot is strange sentence construction that I THINK is coming from the synonym system -- where multiple tenses of a word are considered equivalent -- such as "was, is, are" that kind of stuff. Then keyword/contexts that are tense-sensitive start to sound really silly -- someone says, "Was it fun?" and bot says "It is!" because bot knows "is it fun". And if this bot also knows "was it" but not "was it fun" I think it may pick "is it fun" because it's longer even though it's only a synonym and there's an exact (shorter) match to be had...
True or rumor? And fixable or doomed to be silly?
» More new posts: Doghead's Cosmic Bar