- From: Ian Wilson <ian@emotionai.com>
- Date: Mon, 5 May 2008 12:25:01 +0300
- To: pelachaud@iut.univ-paris8.fr
- Cc: public-xg-emotion@w3.org
Catherine, I think the important element is that a character/system is "doing" something. You mention "ECA's" which tend not to move where perhaps we think moving = doing? For me, my system has action tendancies and is currently focused on what a character would "tend" to do in a virtual world for example. Perhaps the opposite way of thinking is that current research implimentations do not have action tendancies so they are not yet mature, rather than action tendancies being pre mature? Looking to the future it is difficult to imagine an emotion based system that did not output action tendancies I think. Best Regards, Ian -- Ian Wilson CEO Emotion AI www.emotionai.com www.linkedin.com/in/ianwilson/ ian@emotionai.com Company registered in England #6546400 On Mon, 5 May 2008 00:57:17 +0200 (CEST), pelachaud@iut.univ-paris8.fr wrote: > Dear Ian, > >> b. Core 9 Action Tendencies - On looking at this more it seems like a >> "must" candidate. For a system where there is input and output (not >> annotation per se) then that output is very likely to be in the form of >> action tendencies and so this would be a fundamental component I think? > > I agree with you in principles. However it seems to me also that most of > the implementations of emotion model do not deal yet with action tendency. > for example, I am not aware of ECA systems that model action tendency in > the terms of Frijda. > if this item is very premature in research development, shall we include > Action Tendencies in the core element? Isn't a bit to premature? What do > you say? > > Best, > Catherine > >> >> Regards, >> >> Ian >> >> -- >> Ian Wilson >> CEO >> >> Emotion AI >> www.emotionai.com >> www.linkedin.com/in/ianwilson/ >> ian@emotionai.com >> >> Company registered in England #6546400 >> >> >> >> >>
Received on Monday, 5 May 2008 09:25:41 UTC