Adrian Colyer has found a delicious paper (The seven tools of causal inference with reflections on machine learning Pearl, CACM 2018) that works (for me, at least) on multiple levels. Here's the story: " To understand why? and to answer what if? questions, we need some kind of a causal model. In the social sciences and especially epidemiology, a transformative mathematical framework called 'Structural Causal Models' (SCM) has seen widespread adoption." OK. But contemporary machine learning and artificial intelligence are capable only of associative inference. But what about reasoning requiring interventions, and what about predictive and counterfactual reasoning? So, given all this, what would an SCM combining graphical modeling, structural equations, and counterfactual and interventional logic look like? Could we build a machine learning version? And what does this tell us about the structure of contemporary research models?
Today: 5 Total: 100 [Direct link] [Share]
