But right now I'm mostly working on a longer paper about the technical role of emotions and other drives (e.g. hunger) in intelligent control. People used to think these things were chemical and slowish to change just because they were "primitive", but I think there's a reason for why they have been preserved during our evolution. Basically, some things just need longer attention so you have a good chance of completing a task (running away, getting food). If you are distracted too easily, then you waste time going back and forth between projects. But if you aren't distracted easily enough, you get eaten. So you evolve a whole range of different emotions & drives that correspond with particular goals and have duration characteristics appropriate for the typical completion time for the behaviour they motivate.
It's great to think about this stuff and I have had some insight today into how to make my system simpler by adding this in. Of course, adding anything in makes something more complicated, but there may be an important tradeoff here. Still, I'd like to get to really programming and running some models.
The weather here is still great, though a little cooler. The trees are starting to turn colours. Friday I'm attending a big one-day conference on consciousness at the University of Vienna, and then Friday night Will will be here for a week. :-)
Comments