| I just looked through my bookshelves and here’s what I’ve been through since college (in no particular order other than top to bottom on my shelves): 7 books on general ML (highlights: Murphy’s Machine Learning, Hastie et al’s ESL, Koller&Friedman’s PGMs) 5 on more specialized ML (highlights: Agarwal&Chen’s statistical recommender systems book, Manning&Schutze’s statistical NLP, Settles’ active learning) 13 on stats (highlights: Wooldridge’s econometrics of x-section/panel data, Angrist&Pischke’s econometrics) 4 on numerical methods (highlight: Absil et al’s optimization on matrix manifolds) 4 on CS (highlight: CLRS’s intro to algorithms) 10 on calculus/geometry/topology/algebra (highlights: Bachman’s geometric approach to differential forms, Hestenes&Sobzyk's
Clifford algebra to geometric calculus) 8 on fiction writing (highlight: Bickham’s Scene & Structure) And Rosenberg‘s Nonviolent Communication (not a textbook, but still a highlight worth mentioning). It amounts to between 3 and 4 per year. Looking back and counting them up, my reaction is holy crap that’s a lot, but that’s kinda the point. Each year it is a reasonable amount of self study. Not nothing, but not anything crazy. Over the course of many years, it adds up to a hell of a lever. |