I've prefaced a few of my talk recently with some caveats about models. The selection of a model, I argue, presupposes the ouycome of the research in which it is used. We have, as this article notes, a bias toward simle models. This has an impact in the design of thinking achines, and also in our understanding of thinking. Jumping from phenomenon to simple model to prediction is a lot more complex - and potentially misleding - than jumping from phenomenon straaight to prediction.
Today: 4 Total: 103 [Direct link] [Share]
