Learning based on positive/negative feedback

[kevinvr]

I think this is a great idea.  I am constantly rewriting answer lines and trying to "educate" kendy and I think this will make the task easier.

[onthecuttingedge2005]

On my last project with human emotions, I centered the emotional root called trust, all emotions being link to trust, so that even if the bot forgave you it still had mistrust built up, something like what I wrote in the past could help here, If for say the bot mistrusted you, its learning level decreased to Zero. there are many possibilities here.

just passing along some idea's
Jerry

[raybe]

OTCE, Good to hear from you on Ultra Hal again.

You would know better than I but didn't you script a human emotions project using some developed weighting system for all emotions. As you mentioned the more you, for example insulted Ultra Hal the weight would change to a hire number causing Ultra Hal to remember that you insulted Ultra Hal many times over. This would also cause Ultra Hal to react to the way you treated or interacted with Ultra Hal?

raybe

[onthecuttingedge2005]

OTCE, Good to hear from you on Ultra Hal again.

You would know better than I but didn't you script a human emotions project using some developed weighting system for all emotions. As you mentioned the more you, for example insulted Ultra Hal the weight would change to a hire number causing Ultra Hal to remember that you insulted Ultra Hal many times over. This would also cause Ultra Hal to react to the way you treated or interacted with Ultra Hal?

raybe

Yeah, I think it was called the human emotions engine, I had also made a Hap scripted to work with the emotions on the fly, it has been so long ago.

Jerry