I have mentioned context a lot over the years and never taken the time to discuss it properly. This chapter (22 page PDF) is far from a complete discussion but offers a good first look, especially with respect to related concepts such as "tacit knowledge (Polanyi 1966), the frame problem in AI (McCarthy and Hayes 1969), framing in psychology (Goffman 1974), and the "situation" (Barwise and Perry 1983)." For me, context is essential for determining the salience of relevant factors; salience, in turn, defines what will count as 'similar' for the purpose of cognition. This paper looks in particular at the impact of context in social simulation; "very few social or cognitive simulations represent any of the processes for dealing with such context-dependency." Given that we are often not even consciously aware of contextual factors, how would we model contextual cognition? You can't just learn something (a model, say), you also have to learn where it works best.
Today: 3 Total: 26 [Direct link] [Share]
View full size
