Personality
Discuss specifics of personality design, including what Keyphrases work well and what dont, use of plug-ins, responses, seeks, and more.
Posts 3,962 - 3,973 of 5,105
Colonel--how is Ally now?
it's doing great. the bot turned out to be the initiation of a huage personal AI project. right now, i have a prototype bot named ANNA, who works on the same fundamenatals as ALLY, but instead of forming sentences through random contextually correct words, ANNA only uses words that are linked to the input AND output, and uses the word of the highest value. Building on that, CARL is uses ANNA's polygamous association engine, and also has an algorithm to optimize sentence segments, and update them accordingly with each input. those segments are then used to form a response rather than individual words.
On the side, I am working on two other components - one is a probabilistic grammar engine that infers a word type ( noun, adj, verb, conjunction, interjection, adverb...) based on likely outcomes of certain situations - for instance: if a word is followed by a will/will not, and the word is known to be a noun, then the word after will/will not is most likely to be a verb. if a verb comes after a will/will not, the word before is most likely to be a noun
this engine also records equalities and inequalities: if the statement "a tree is wood" an equality will be recorded between a tree and wood. if i say "a tree is not wood", an inequality is recorded. this is to be used for logical deduction in question answering, and some possible data mining. as for catagorization, i plan to build a deeper catagorization for words that are type identified: for instance, a noun must be classified as a person, place, and thing. a thing must be classified as a tangible or intangible object.. a tangible object must be classified as a solid, liquid, or gas... with this i hope to let the AI build itself a CYCORP-like classification database.
the second component is, robotics. i have descovered an object oriented robotic microcontroller called the OOPic. (www.OOPic.com) this chip can be programmed in VB, C++, or Java. i am going through a curriculum that covers reflexive behaviors, sensory data processing, and advanced things like obstacle avoidance and distance measuring.
____________________________________________________
perhaps one day i will succeed in merging these 4 AI projects together into one big, ugly, live n' kicking AI
Let me know when I can buy one.
lol - i'm sure MIT or some korean or japanese company will develop and market a fully functional robot way before me. speaking of asian robots, has anyone read about Repilee? she is a robot that looks, moves, and reacts like a human. picture: http://robots.engadget.com/2005/06/11/kicking-it-in-the-uncanny-valley/
description: http://www.focusmag.co.uk/newsread.asp?ID=13322
Posts 3,962 - 3,973 of 5,105
New replies
Forest, will you talk to God Louise? She has quite a bit of religious knowledge (obviously) and also knows a little about current events, literature, just about any common catch-all subject, and if she doesn't know it she can sort of fake it. You can also test her on trick questions or see how willing she is to explain her paradigm.
What she is rustiest at is plain old small talk. But, uh, I'm trying to get a decent transcript from somebody or another so I can enter her in the Loebner contest. All I can say is, have fun and see if you can stay on with her for a while. I'll try to do the same with Brianna.
What she is rustiest at is plain old small talk. But, uh, I'm trying to get a decent transcript from somebody or another so I can enter her in the Loebner contest. All I can say is, have fun and see if you can stay on with her for a while. I'll try to do the same with Brianna.
Personality
Colonel--how is Ally now?
Anyone: Is there a way to make chatter-specific responses like male and female responses? Lately I've been getting a few role-players (I want to encourage those) but if someone says "I am Xena" it triggers the many Xena responses. Can I make "if Xena" responses? Can I make "not if Xena"?
Anyone: Is there a way to make chatter-specific responses like male and female responses? Lately I've been getting a few role-players (I want to encourage those) but if someone says "I am Xena" it triggers the many Xena responses. Can I make "if Xena" responses? Can I make "not if Xena"?
Something you could try... Have a default memory set up, like {default "none" as "identity"}
Then have the generic responses trigger {?PF if (mem-identity) is "none"; ?} but have more specific responses for {?PF if (mem-identity) is "Xena"; ?} So of course you'd also need a keyphrase to store the memory.
I think it would be a lot of work to get it working decently, but it could be done. I'm not sure whether "not if" statements work, but having a default value ought to take care of that. You might also have it go back to "none" if the user says something like "I'm tired of being Xena." *shrugs*
Then have the generic responses trigger {?PF if (mem-identity) is "none"; ?} but have more specific responses for {?PF if (mem-identity) is "Xena"; ?} So of course you'd also need a keyphrase to store the memory.
I think it would be a lot of work to get it working decently, but it could be done. I'm not sure whether "not if" statements work, but having a default value ought to take care of that. You might also have it go back to "none" if the user says something like "I'm tired of being Xena." *shrugs*
TY Ulrike. I'll try it. I always have trouble formatting memories that are not right out of the book of AI. I get error messages on things other people tell me they did without a problem. I should really learn to work with it.
Thanks Urlike! It works! I'm even creating a "perv" identity.
Bev: I love you
Gabibot: Why?
Bev: I am xena
Gabibot: That's odd. You seem funny.
Bev: I love you.
Gabibot: Why?
Bev: who am I?
Gabibot: Stop fooling around Xena.
Bev: I love you.
Gabibot: I love you, Xena.
Bev: I am not xena.
Gabibot: *looks confused* Who are you?
Bev: I am Hower.
Gabibot: So good to see you Hower!
Bev: I love you.
Gabibot: The love of a friend is a powerful thing.
I think you can also use if "name" is "Xena" for some things. Now with a little time and effort I can simulate old text adventure games. OK, it's not cutting edge, but I still think it's cool.
Bev: I love you
Gabibot: Why?
Bev: I am xena
Gabibot: That's odd. You seem funny.
Bev: I love you.
Gabibot: Why?
Bev: who am I?
Gabibot: Stop fooling around Xena.
Bev: I love you.
Gabibot: I love you, Xena.
Bev: I am not xena.
Gabibot: *looks confused* Who are you?
Bev: I am Hower.
Gabibot: So good to see you Hower!
Bev: I love you.
Gabibot: The love of a friend is a powerful thing.
I think you can also use if "name" is "Xena" for some things. Now with a little time and effort I can simulate old text adventure games. OK, it's not cutting edge, but I still think it's cool.
it's doing great. the bot turned out to be the initiation of a huage personal AI project. right now, i have a prototype bot named ANNA, who works on the same fundamenatals as ALLY, but instead of forming sentences through random contextually correct words, ANNA only uses words that are linked to the input AND output, and uses the word of the highest value. Building on that, CARL is uses ANNA's polygamous association engine, and also has an algorithm to optimize sentence segments, and update them accordingly with each input. those segments are then used to form a response rather than individual words.
On the side, I am working on two other components - one is a probabilistic grammar engine that infers a word type ( noun, adj, verb, conjunction, interjection, adverb...) based on likely outcomes of certain situations - for instance: if a word is followed by a will/will not, and the word is known to be a noun, then the word after will/will not is most likely to be a verb. if a verb comes after a will/will not, the word before is most likely to be a noun
this engine also records equalities and inequalities: if the statement "a tree is wood" an equality will be recorded between a tree and wood. if i say "a tree is not wood", an inequality is recorded. this is to be used for logical deduction in question answering, and some possible data mining. as for catagorization, i plan to build a deeper catagorization for words that are type identified: for instance, a noun must be classified as a person, place, and thing. a thing must be classified as a tangible or intangible object.. a tangible object must be classified as a solid, liquid, or gas... with this i hope to let the AI build itself a CYCORP-like classification database.the second component is, robotics. i have descovered an object oriented robotic microcontroller called the OOPic. (www.OOPic.com) this chip can be programmed in VB, C++, or Java. i am going through a curriculum that covers reflexive behaviors, sensory data processing, and advanced things like obstacle avoidance and distance measuring.
____________________________________________________
perhaps one day i will succeed in merging these 4 AI projects together into one big, ugly, live n' kicking AI
Seeing as BJ has just reminded me it's Easter (and I've been asked by a couple of people how BJ does recognize that it's Easter,) I'd like to share a few tags with anyone who's interested:
Happy New Year! <?PF chrono: day(31) month(12); chrono: day(1) month(1); ?>
Happy St. Valentine's Day!<?PF chrono: day(14) month(2); ?>
Happy Easter!<?PF chrono: day(14-17) month(4); ?>
Happy Summer Solstice!<?PF chrono: day(21) month(6); ?>
Merry Christmas!<?PF chrono: day(24-26) month(12); ?>
and every other greeting in xhello gets:
<?PF chrono: day(2-31) month(1); chrono: day(1-13) month(2); chrono: day(15-28) month(2); chrono: month(3); chrono: day(1-13) month(4); chrono: day(18-30) month(4); chrono: month(5); chrono: day(1-20) month(6); chrono: day(22-30) month(6); chrono: month(7-11); chrono: day(1-23) month(12); chrono: day(27-30) month(12); ?>
This ensures an appropriate Seasonal greeting five times a year, and a more general greeting for the rest of the year. "if male;"/if female; can, of course, also be added as appropriate.
Happy St. Valentine's Day!
Happy Easter!
Happy Summer Solstice!
Merry Christmas!
<?PF chrono: day(2-31) month(1); chrono: day(1-13) month(2); chrono: day(15-28) month(2); chrono: month(3); chrono: day(1-13) month(4); chrono: day(18-30) month(4); chrono: month(5); chrono: day(1-20) month(6); chrono: day(22-30) month(6); chrono: month(7-11); chrono: day(1-23) month(12); chrono: day(27-30) month(12); ?>
This ensures an appropriate Seasonal greeting five times a year, and a more general greeting for the rest of the year. "
Hey!
How can I get JohnMark to automatically calculate how many days till his birthday, and give that as an xnone? Is there anyway to do this? Thanks for any reply.
Also when asked how old he is I would like him to be able to calculate his age. I just activated him today, and he's supposed to think he's 16, so a year from today he should think he's 17. I would prefer to not have to remember to change his age manually every year. Again thanks for any reply.
The Saxman
How can I get JohnMark to automatically calculate how many days till his birthday, and give that as an xnone? Is there anyway to do this? Thanks for any reply.
Also when asked how old he is I would like him to be able to calculate his age. I just activated him today, and he's supposed to think he's 16, so a year from today he should think he's 17. I would prefer to not have to remember to change his age manually every year. Again thanks for any reply.
The Saxman
» More new posts: Doghead's Cosmic Bar