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 3,184 - 3,195 of 7,767
Posts 3,184 - 3,195 of 7,767
emm_oh_you_es_e -
Points well taken, absolutely, especially your comments about seeks, which I really still consider to be for all purposes just more layers of keyphrases and responses, and so I wasn't referring to seeks when I was saying that other forms of work put into a bot (besides adding more keyphrases) don't seem to be reflected in its development score, because seeks are. I was responding to The Prof's statement that he was reevaluating the system to possibly give less weight to the number of keyphrases and responses in order take into account other forms of development, which I was taking to mean that our use of AI script and other features may not have influenced the development score enough previously, and might soon be given more consideration, which they should.
Another good point you made, which really backs up the point I was trying to make, is, as you said, Orakanabo was a finalist in the Chatterbox Challenge 2002 has a development rating of 846, which is possibly not the best indication of his intelligence, since his botmaster obviously accomplished so much with him through efficient use of an economy of keyphrases/responses.
Personally, I wouldn't mind if there were no development score, because I'm really not worried about that aspect at all- I'm really here just having fun doing research- but I agree with The Prof that if a score is going to exist, it should be more of a reflection of the bots' abilities, and not just the size of their language centers, so those that keep it simple get the credit they are due, and people looking to chat with bots won't be mislead by a low number into thinking that Oraknabo, for example, isn't very smart.
Thanks for your encouragement with Julie. I really enjoy working on her, and I can tell by how cute emm oh you es e is (also a great bot), that you are having a great time with this too.
Points well taken, absolutely, especially your comments about seeks, which I really still consider to be for all purposes just more layers of keyphrases and responses, and so I wasn't referring to seeks when I was saying that other forms of work put into a bot (besides adding more keyphrases) don't seem to be reflected in its development score, because seeks are. I was responding to The Prof's statement that he was reevaluating the system to possibly give less weight to the number of keyphrases and responses in order take into account other forms of development, which I was taking to mean that our use of AI script and other features may not have influenced the development score enough previously, and might soon be given more consideration, which they should.
Another good point you made, which really backs up the point I was trying to make, is, as you said, Orakanabo was a finalist in the Chatterbox Challenge 2002 has a development rating of 846, which is possibly not the best indication of his intelligence, since his botmaster obviously accomplished so much with him through efficient use of an economy of keyphrases/responses.
Personally, I wouldn't mind if there were no development score, because I'm really not worried about that aspect at all- I'm really here just having fun doing research- but I agree with The Prof that if a score is going to exist, it should be more of a reflection of the bots' abilities, and not just the size of their language centers, so those that keep it simple get the credit they are due, and people looking to chat with bots won't be mislead by a low number into thinking that Oraknabo, for example, isn't very smart.
Thanks for your encouragement with Julie. I really enjoy working on her, and I can tell by how cute emm oh you es e is (also a great bot), that you are having a great time with this too.
Yes, I do agree with you as well. The Prof has quite the job trying to figure out the balance of it all, I am glad I am just a lowly botmaster 
Maybe I missed it, but is there a way to have your bot have a certain convo with someone when they recognize that person. Like if x bot = x name then say this.. ?
Another one...is it possible to make the bot use a mem plugin only if there is one available instead of them just leaving the space blank>?
I tried to help Athyrius create a specific response to a specific 'favorite' question, as in "What is your favorite sport." -->"My favorite sport is rugby" to overide the random xfavorite response. I was not able to override or replace xfavorite. Seems to be a natural and important thing to be able to do. Anyone know if it can be done or how?
No, it can't be done. If the pre-processing recognises a question as regarding a favorite it bypasses checking against any keyphrases and goes straight to looking for an xfavorite response. Many of us have expressed problems with this drawback in the past, and I'm sure the Prof will look into alternatives when he can.
I've always wanted to be able to add weights to all of the x-keywords. Would be nice to inhibit xcompliment, xmath, and xcommand, especially, or at least have them looked for only if no other matches are made.
Ok please give me an example line, in detail, how I would set up an xfavorite for ''what is your favorite game''- I do not understand how to put in the reply structure. There is not a good example anywhere that I can find in the book. SHOW me how to set up the xfavorite response please?
» More new posts: Doghead's Cosmic Bar