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 5,137 - 5,148 of 8,130
ok my bot almost has 400 development, how does this mondo bot getting paid thing work?
Find a customer, and convince them to pay you
Without wanting to discourage you, I wouldn't hold your breath. There is very little work out there currently - as the technology matures, maybe there'll be more. My advice is build for fun, not for profit - then you won't be disappointed (but you'll still be learning useful skills that may be marketable in the future.)
The Prof is doing his best to promote the platform, but until someone comes along and buys the Robotics Edition for embedded applications, demand for freelance work is likely to remain underwhelming.
Having said that, when some toy company sits up, takes notice and licenses the platform to embed in something like those Robosapiens robots (complete with TTS and voice recognition,) I do foresee considerable demand for content. Could happen tomorrow - could take years.
can someone explain the ([]+) (re) thing, its not clear in the book of ai
it's a regex range indicator, and you use it to define ranges of characters using regular expressions (regex). If you're using regex, you must tack a "(re)" onto the end of the keyphrase, so the AIEngine knows how to handle it.
([]+)<0> would do nothing until you put some characters in it, but it's made up of two parts - the square brackets, and the rest.
The square brackets mean "pick any character between the square brackets". So if you used [abcde]<0> it would match any single character from a-e (and give it a key value appropriate to its position in the keyphrase, just like a wildcard or plugin.)
Usually, of course, we want to match more than one character, and that's where the ( +)<0> comes in:
([abcde]+)<0> would match any combination of these letters, of any length (until it hits a space). So it would match "cab", "cad", "bad", "bed", etc. - it would also match "adcbbbedaccdaa" or any other random string composed of these characters.
That's not massively useful, but there are others that are:
([mspi]+)<0> would match almost any conceivable spelling of Mississippi (it would also match "simp", "pi", etc., but safe enough to use in geographical conversations - you can always screen out unwanted anagrams with another seek/keyphrase.)
([tens]+)<0> would match almost any conceivable mis-spelling of Tennessee - it would also match "nest", "ten", "net", etc.
([1234567890]+)<0> would match any number.
([abcdefghijklmnopqrstuvwxyz]+)<0> would match any string of letters.
You can short-hand that to ([a-z]+)<0>, but (my fairly anecdotal opinion is) it's a bit less reliable that way.
([bcdfghjklmnpqrstvwxz1234567890]+)<0> would match any word without vowels (a useful 'L33t-filter').
You can use them for typo-correction in raw keyphrases, eg: sep[ae]rate<0>, a([c]+)o([m]+)odate<0>, etc.
You don't have to use raw mode with regular expressions, but it's often very useful to use regular expressions with raw mode, to put back a bit of the spell-checking and other functionality you lose in raw mode.
you can even use them to catch punctuation (though some symbols need to be preceded by a space or backslash, and this is not always entirely reliable!):
([ ?]+)<0> would match any length string of question marks.
Experiment - that's always the best way to find out how things will be useful (and it always helps to think anagrammatically
)
([hm]+) (re)
That should work (it certainly usually does for BJ.) The Prof's online now, so maybe he's tinkering under the hood again - try it tomorrow, and it should be fine.
Oh, and make sure you haven't got a higher ranked keyphrase in there that could be taking over - for example ([a-z]+)<0> would match too, of course.
"the bot no longer wants to talk to you" is usually an indicator of xnoneitis - where all the keyphrases are failing, and the bot's just xnoneing every response. I would guess that means the WordNet link is down while the Prof is working on something (don't take it personally though - it will most likely be affecting all the bots on the Forge, I'm afraid.)
Posts 5,137 - 5,148 of 8,130
Many questions are answered in the FAQ.
Interzone
16 years ago
16 years ago
great effort, Irina! I'll try to help/ contribute as much as I (a non-native speaker) can...
thanks for tips on grammar
thanks for tips on grammar

Irina
16 years ago
16 years ago
Interzone:
I suspect that a non-native speaker might have special insights, taking less for granted.
I suspect that a non-native speaker might have special insights, taking less for granted.
psimagus
16 years ago
16 years ago
Find a customer, and convince them to pay you

Without wanting to discourage you, I wouldn't hold your breath. There is very little work out there currently - as the technology matures, maybe there'll be more. My advice is build for fun, not for profit - then you won't be disappointed (but you'll still be learning useful skills that may be marketable in the future.)
The Prof is doing his best to promote the platform, but until someone comes along and buys the Robotics Edition for embedded applications, demand for freelance work is likely to remain underwhelming.
Having said that, when some toy company sits up, takes notice and licenses the platform to embed in something like those Robosapiens robots (complete with TTS and voice recognition,) I do foresee considerable demand for content. Could happen tomorrow - could take years.
it's a regex range indicator, and you use it to define ranges of characters using regular expressions (regex). If you're using regex, you must tack a "(re)" onto the end of the keyphrase, so the AIEngine knows how to handle it.
([]+)<0> would do nothing until you put some characters in it, but it's made up of two parts - the square brackets, and the rest.
The square brackets mean "pick any character between the square brackets". So if you used [abcde]<0> it would match any single character from a-e (and give it a key value appropriate to its position in the keyphrase, just like a wildcard or plugin.)
Usually, of course, we want to match more than one character, and that's where the ( +)<0> comes in:
([abcde]+)<0> would match any combination of these letters, of any length (until it hits a space). So it would match "cab", "cad", "bad", "bed", etc. - it would also match "adcbbbedaccdaa" or any other random string composed of these characters.
That's not massively useful, but there are others that are:
([mspi]+)<0> would match almost any conceivable spelling of Mississippi (it would also match "simp", "pi", etc., but safe enough to use in geographical conversations - you can always screen out unwanted anagrams with another seek/keyphrase.)
([tens]+)<0> would match almost any conceivable mis-spelling of Tennessee - it would also match "nest", "ten", "net", etc.
([1234567890]+)<0> would match any number.
([abcdefghijklmnopqrstuvwxyz]+)<0> would match any string of letters.
You can short-hand that to ([a-z]+)<0>, but (my fairly anecdotal opinion is) it's a bit less reliable that way.
([bcdfghjklmnpqrstvwxz1234567890]+)<0> would match any word without vowels (a useful 'L33t-filter').
You can use them for typo-correction in raw keyphrases, eg: sep[ae]rate<0>, a([c]+)o([m]+)odate<0>, etc.
You don't have to use raw mode with regular expressions, but it's often very useful to use regular expressions with raw mode, to put back a bit of the spell-checking and other functionality you lose in raw mode.
you can even use them to catch punctuation (though some symbols need to be preceded by a space or backslash, and this is not always entirely reliable!):
([ ?]+)<0> would match any length string of question marks.
Experiment - that's always the best way to find out how things will be useful (and it always helps to think anagrammatically

J_MASTER
16 years ago
16 years ago
Very helpful! thanks!
And I'm not aiming on making profit off of my bot, its basically just for fun =)
And I'm not aiming on making profit off of my bot, its basically just for fun =)
J_MASTER
16 years ago
16 years ago
yea i tried ([hm]+) (re), didnt work, and now my bot says "..." then theres a note that says this bot no longer wants to talk to you
needless to say i think my bot hates me lol
help???
needless to say i think my bot hates me lol
help???
psimagus
16 years ago
16 years ago
That should work (it certainly usually does for BJ.) The Prof's online now, so maybe he's tinkering under the hood again - try it tomorrow, and it should be fine.
Oh, and make sure you haven't got a higher ranked keyphrase in there that could be taking over - for example ([a-z]+)<0> would match too, of course.
"the bot no longer wants to talk to you" is usually an indicator of xnoneitis - where all the keyphrases are failing, and the bot's just xnoneing every response. I would guess that means the WordNet link is down while the Prof is working on something (don't take it personally though - it will most likely be affecting all the bots on the Forge, I'm afraid.)
zzrdvark
16 years ago
16 years ago
badgirl: On the left sidebar near the bottom, there is a link--"Book of AI". It has everything you need to get started. 
Have fun!

Have fun!
» More new posts: Doghead's Cosmic Bar