Thanks for the suggestion! Most of the relevant knowledge from Causality can be picked up in the recommended chapters of Probabilistic Graphical Models. Pearl’s textbook will still get you some additional knowledge, but I’m wary of putting it as a prereq for fear of draining resources (Pearl can be a bit dense) without too much added benefit over the PGM book (which is in the “basics” section, because causal models are important in pretty much all of the subfields, not just DT.)
That’s interesting. It’s hard for me to tell what’s easy anymore (Pinker’s curse of knowledge) in this area. I do think Pearl’s book isn’t very hard compared to a lot of other math books, though. For example, I find “Unified Methods for Censored Longitudinal Data and Causality” a ton harder to read.
Naturally, I have a lot of arguments for reading Pearl, but my strongest argument against reading Pearl is that he doesn’t tackle statistical issues at all, and they are very important in practice.
For the Decision Theory section, may I suggest a reference to a causality text, like Pearl 1999?
Thanks for the suggestion! Most of the relevant knowledge from Causality can be picked up in the recommended chapters of Probabilistic Graphical Models. Pearl’s textbook will still get you some additional knowledge, but I’m wary of putting it as a prereq for fear of draining resources (Pearl can be a bit dense) without too much added benefit over the PGM book (which is in the “basics” section, because causal models are important in pretty much all of the subfields, not just DT.)
That’s interesting. It’s hard for me to tell what’s easy anymore (Pinker’s curse of knowledge) in this area. I do think Pearl’s book isn’t very hard compared to a lot of other math books, though. For example, I find “Unified Methods for Censored Longitudinal Data and Causality” a ton harder to read.
Naturally, I have a lot of arguments for reading Pearl, but my strongest argument against reading Pearl is that he doesn’t tackle statistical issues at all, and they are very important in practice.
Ah, makes sense.