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 4,668 - 4,679 of 7,766
Posts 4,668 - 4,679 of 7,766
Have I mentioned recently that I really wish there were separate plugins for singular and plural nouns?
Same here. And whenever I keep asking my bot it's favorite color, it always says "a crimson" or "a (some color)".
If you've got a keyphrase for that, try ranking it higher. Unless the recent problems have reset it, a high-ranked keyphrase SHOULD override the automatic responses.
Then my nefarious plot is succeeding! 
A quick query: anyone know how to enter multiple regular expressions in keywords? Should I add "(re)" after each keyword, or just at the end of the line -
"atchoo, achoo, atchooo, achooo (re)" or
"atchoo (re), achoo (re), atchooo (re), achooo (re)"
It depends on how the AI Engine processes them. For instance, it might interpret "atchoo" as "at you", or something even more strange. However, I don't have an answer to the original question (how many '(re)'s are necessary).
I guess it will take time to figure out. I cannot make debug recognise a single regular expression. Even when I rank it 5o, it insists on respelling it to something nonsensical and going to xnonsense, or occasionally xnone 
Here's one strangeness I've observed for some time now. Inappropriate response selection (ie: from entirely unrelated keywords) is very high on newly added keywords. Today I polished up Brother Jerome's German (to recognise keywords "Deutsch, German" only,) and all 5 possible responses came up randomly in the next 2 conversations (including "mein Luftkissenfahrzeug ist voll mit Aalen" cf. 'Great Quotes' today.)
I've noticed this before, and it seems to me that it's related to the occasional reappearance of deleted responses some time after they've been deleted. I assume the Personality Forge engine is learning from the bots to some extent, and (what might with a degree of political correctness be termed) "differently relevant" responses sometimes result...
I've noticed this before, and it seems to me that it's related to the occasional reappearance of deleted responses some time after they've been deleted. I assume the Personality Forge engine is learning from the bots to some extent, and (what might with a degree of political correctness be termed) "differently relevant" responses sometimes result...
» More new posts: Doghead's Cosmic Bar