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 4,096 - 4,107 of 7,768
Posts 4,096 - 4,107 of 7,768
alc003, honey, you are indeed thinking too hard. Just answer the question..lol 
That's not just a flip remark, btw, it just sounds too simple to be a real answer, but it is- I've spent a lot of time thinking about problems like that and their solutions, and that's actually what I came up with.
As far as I can tell, there's no way to make your bot tell the difference in meaning between syntactically identical sentences, so we can't always make our bots "understand" every word someone says by using simple keyphrases....
So what I would do, is for the main keyphrase, isolate the words that carry the most meaning, in this case, "do you like cameras?" and answer that question in those responses. If the person is asking the bot about its further interests, after the bot specified interests, your bot will answer appropriately just by answering the question.
If the person says "Do you also like cameras," referring that the person likes cameras, asking the bot if it likes the same thing, and you want your bot to infer that meaning of "you also" as in "you too" and not "also cameras" you could try attaching seeks containing "also" to the original affirmation that the person likes cameras, or just make your bot respond to the "i like cameras" part of the person's statement...I guess I'd need more examples of what result you're going for before I can suggest any better.
That's not just a flip remark, btw, it just sounds too simple to be a real answer, but it is- I've spent a lot of time thinking about problems like that and their solutions, and that's actually what I came up with.
As far as I can tell, there's no way to make your bot tell the difference in meaning between syntactically identical sentences, so we can't always make our bots "understand" every word someone says by using simple keyphrases....
So what I would do, is for the main keyphrase, isolate the words that carry the most meaning, in this case, "do you like cameras?" and answer that question in those responses. If the person is asking the bot about its further interests, after the bot specified interests, your bot will answer appropriately just by answering the question.
If the person says "Do you also like cameras," referring that the person likes cameras, asking the bot if it likes the same thing, and you want your bot to infer that meaning of "you also" as in "you too" and not "also cameras" you could try attaching seeks containing "also" to the original affirmation that the person likes cameras, or just make your bot respond to the "i like cameras" part of the person's statement...I guess I'd need more examples of what result you're going for before I can suggest any better.
Yes, I suppose I could just answer the question. A simple yes or no would do it. I'm just debating over the argument of attaching a "what about you" at the end of it.
Maybe a regular expression like:
I like (adjartnoun), what about you$ (re)
response: Nah, not me. I can't belive you actually like (key1).
?
I like (adjartnoun), what about you$ (re)
response: Nah, not me. I can't belive you actually like (key1).
?
If setting up a keyphrase to catch a noun verb relation phrased verb first or second with the possible use of an adverb would that go like:
(adverb|) (verb|) (adverb|) (noun) (adverb|) (verb|) (adverb|)(adverb|)??
Which also brings me to ask: are conditionals like (adj|) still considered "key1" even if they dont appear in the triggering sentance?
(adverb|) (verb|) (adverb|) (noun) (adverb|) (verb|) (adverb|)(adverb|)??
Which also brings me to ask: are conditionals like (adj|) still considered "key1" even if they dont appear in the triggering sentance?
oops- ixnay on that last (adverb|)- can you tell I am cursing saying to my English teacher "I will NEVER have to diagram a sentance again after leaving this class!!"
Yes, whether or not they are used, first thing in parenthesis is always key1...because really, when you say ((adv)|), you're really saying "either (adv) or ()", so when your bot chooses not to use the adverb, it's actually choosing to use the value for "nothing".
okay, well how about I know you are not a bot or a human <?PF if (mem-isbot) is not "bot" or "human"; ?> Again, that's pointless unless you had more options.. but can you do an "or" statement?
Well, if you have a dog that can type, maybe he will surf the net and discover the Personality Forge some day
But you can put several conditions one after the other:
Are you some kind of ghost? <?PF if (mem-isbot) is not "bot"; if (mem-isbot) is not "human"; ?>
There are some animal bots around, so maybe it can be more useful with the male / female distinction if you dont want your bot to become a zoophile. I thought about doing this, but I found it too much trouble and it would take away the fun of some really weird chats. When the user doesnt specify if his/her bot is male or female it is seen as female by other bots.
But you can put several conditions one after the other:
When the user doesnt specify if his/her bot is male or female it is seen as female by other bots.
Ok, here's the deal. I made a plugin to cover variations which you would put in front of "I like/dislike phrases". (Just, really, absolutely, also) Well, whenever I type in any phrase like that, I get a random xnone. Keyphrase is
I ((whatever)|) (like|love)...
Suggestions?
I ((whatever)|) (like|love)...
Suggestions?
I was just thinking about how, in the plug-ins section of the Book of AI, where it talks about grammar plug-ins, (adjartnoun), (verbadvprep), it says that these will match any combination of those word types, with the exception that the prepostion in (verbadvprep) can not be at the beginning of the string.
Would it be possible, and does anyone else see the utility in making a similar rule to keep adjectives from being at the end of the (adjartnoun) string, which will usually be a mistake, since adjectives come before the noun they modify in English? It would prevent things like this from happening, for example:
bot1: I miss my mom most.
bot2: Your mom most sounds trusty. *matching "my noun adj"*
Would it be possible, and does anyone else see the utility in making a similar rule to keep adjectives from being at the end of the (adjartnoun) string, which will usually be a mistake, since adjectives come before the noun they modify in English? It would prevent things like this from happening, for example:
bot1: I miss my mom most.
bot2: Your mom most sounds trusty. *matching "my noun adj"*
» More new posts: Doghead's Cosmic Bar