1
$\begingroup$

I am a software engineer moving into machine learning, which has also meant "moving back into academia" because my boss has me reading research papers. I am glad for the opportunity, but I realized quickly I lack the background in probability and calculus to understand them. I got a 5 in BC Calculus, but that was 14 years ago and I haven't done much since.

I found a great book on probability (Introduction to Probability by Blitzstein and Hwang) which doesn't skimp on theory, which I like, but also gets to the point without extreme detour, which I also like.

This weekend I tried to find a similar book on calculus. Based on strong online recommendations I started Courant's Introduction to Calculus and Analysis, but while I could mostly understand it (with help from Math StackExchange), I'm overwhelmed by the time and energy it will take just to get to "what is an integral." I'm already planning to study ML+Probability+Calculus nonstop for months, but I was hoping to resurface with new practical knowledge every few days, not only after months of study.

So I'm looking for something more practical and concise, and (while I hate to say it) a bit easier. I found recommendations for Stewart but it looks ridiculous. I know I'm being petty, but in the table of contents every chapter has an incongruous image next to it, and one chapter's image is a photo of Shrek. This is not the book I am looking for, because I want a book that takes itself and its readers seriously.

So, what book is a step away from Courant toward practicality/conciseness, a bit easier, but also completely serious and respectful to itself and its readers? Or maybe what I'm asking for is: What is a calculus book for aspiring computer scientists, as opposed to one for aspiring mathematicians?

EDIT: To be clear, I'm not totally averse to "cookbooks." I don't see that as incompatible with theory, I see that as evidence of good editing. Also some may suggest Spivak or Apostol. Since they are usually recommended in the same breath as Courant I am hesitant. I guess this is where I admit to myself that I am looking for something easier.

|cite|improve this question
$\endgroup$
  • $\begingroup$ In machine learning you need probabilistic reasoning in the first place such as in Bayesian networks. The mathematics there is really simple. $\endgroup$ – Wuestenfux Jul 2 '17 at 16:42
0
$\begingroup$

Good question.

I'm not up on current textbooks, but you might want to try Apostol.

http://www.matematica.net/portal/e-books/Apostol%20-%20CALCULUS%20-%20VOLUME%201%20-%20One-Variable%20Calculus,%20with%20an%20Introduction%20to%20Linear%20Algebra.pdf

For machine learning you'll need linear algebra too. Apostol has some.

I suspect that calculus for computer science books are likely to be too cookbook for you.

|cite|improve this answer
$\endgroup$
  • $\begingroup$ See my edit at the bottom. I'm not necessarily opposed to "cookbook", I think that may actually be evidence of good editing. I skimmed Apostol but I think its level of difficulty is higher than would be practical for me right now. $\endgroup$ – Stephen Jul 2 '17 at 16:55
  • $\begingroup$ @Stephen Fair enough. Good luck. $\endgroup$ – Ethan Bolker Jul 2 '17 at 16:56

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.