Live data from Hacker News

Mathematics for the Adventurous Self-Learner

neilwithdata.com

181–190 of 216 posts

Re: Mathematics for the Adventurous Self-Learner

#181

Does anybody have any experience with How to Prove It? by Velleman? Recently I was thinking of starting on it, but I'm not sure about the level of commitment necessary.

I worked through this book to learn how to do proofs. It turned out way more fun than I expected. The book really did demystify proofs for me. It took several months of studying - there are many exercises. But completely worth it. I'm glad I have read this book before studying Group theory and Real analysis.

Re: Mathematics for the Adventurous Self-Learner

#182
post #48

Earlier quoted context omitted.

So, you've made all that effort, how does it help you in your new role as a data scientist? Is there anything you do now that requires "mathematical maturity"? Or is it something that can be learned much quicker on as needed basis?

There are a lot of charlatans in the Data Science space who lack the necessary mathematical background for their roles. For me it was necessary to get a rigorous understanding of probability theory, and applied probability theory is basically what mathematical statistics is about. My background was CS and software so R, Python, data visualisation and ML operationalization is by and large the easy part of Data Science…

I do deep learning research for a living. I've taken graduate classes in probability, stochastic processes, optimization algorithms, and signal analysis (ECE PhD). I almost never completely understand what's going on under the hood of my models as soon as they get larger than a single neuron XOR mapper. That does not prevent me from finding ways to improve the performance of very large models (millions of parameters and dozens of layers). I agree that there are some papers (or the two books you mentioned) that can be quite dense and heavy on math, but I can't say I've ever felt like I needed any math other than basic calculus, linear algebra, and prob/stats 101 to understand almost all ML methods that people actually use in real world. Obviously if you want to make breakthroughs in theoretical ML, then sure, you do need the mathematical maturity (mostly because you will need to be formally proving things), but if you're a regular data scientist? Can you give some example what kind of math is involved in your predictive models?

Re: Mathematics for the Adventurous Self-Learner

#183

Do any of you all have some tips for understanding mathematical notation? I feel this is often poorly explained, and it feels like a language all its own that just does not speak to me. I did pretty well in calculus, but I still don't really understand what the dx was supposed to represent and in reality I was just really good at pattern matching when it wasn't supposed to be there anymore. I try to read papers now a…

Sounds like you might simply not understand the definitions for these operators and symbols. In other words, it's not a notation problem. I find that it's helpful to mentally replace the symbols like dy/dx, sum, lim, integral, and so on with the concepts they represent. That is, go from operators to definitions.

Re: Mathematics for the Adventurous Self-Learner

#184
Another book similar to Morris Kline's _Mathematics for the Nonmathematician_, which the OP mentioned, is Lancelot Hogben's _Mathematics for the Million_. Originally published in 1937, it has been in print ever since, through several revisions. This also takes a historical approach, beginning with numbers and counting, measure and Greek geometry; and eventually covering calculus, matrices, probability, and statistics.

You can take a look at it, at the Internet Archive,

https://archive.org/details/HogbenMathematicsForTheMillion/m...

Re: Mathematics for the Adventurous Self-Learner

#185
post #169

Earlier quoted context omitted.

That is extremely inspiring. Can you speak more to how you went from industry to a community college to getting into and completing a PhD program? Did you quit your job to return back to school? What area did you end up specializing in and what do you do now?

Long and short, I got sick of the tedium in web development, quit my job and went back to school. The dot-com bubble had just burst, and I had been taking occasional classes including a very inspirational data structures course which planted the seed with formal proofs. After I went back to school, I tutored in the math study center to pay the bills, which really helped cement not just the learning but also the notio…

> had I known better I'd have gone into computer science because that's where I see the most progress being made.

Interestingly this is similar to what my two advisors (one from the math department and one from CS) suggested to me. It would be easier to do the math I like in a CS department than it would be to do the CS I like in a math department. Do you feel like math departments are more conservative when it comes to working outside the discipline?

Re: Mathematics for the Adventurous Self-Learner

#186
post #18

Earlier quoted context omitted.

Yep, brilliant.org + the first book from minireference.com are what OP needs.

I found that the community on brilliant.org can be quite toxic. I was a member for awhile, but quit due to the people on the site.

It gives me great pain to hear this.

Please do email me silas [at] brilliant.org, so I can figure out which failure mode you hit and evaluate if it has improved at all since you quit.

In addition, as others have noted, our courses have no community aspect to them at all.

Re: Mathematics for the Adventurous Self-Learner

#187

Do any of you all have some tips for understanding mathematical notation? I feel this is often poorly explained, and it feels like a language all its own that just does not speak to me. I did pretty well in calculus, but I still don't really understand what the dx was supposed to represent and in reality I was just really good at pattern matching when it wasn't supposed to be there anymore. I try to read papers now a…

I think usually the problem is "almost getting it" and trying to move forward, which means small uncertainties add up and all the sudden one is totally lost without being sure exactly why. So it's important to go back and make sure each piece of notation is crystal clear before moving forward.

Any statement in math is meant to be directly translatable to human language. You should be able to read it out loud in English and know exactly what you mean when you say it.

Unfortunately, sometimes math uses awful notation. For example, df/dx. This is a case where df doesn't mean anything (or at least it's not normally well-defined), and dx doesn't mean anything either (same comment). But the notation as a whole means something. If we write g = df/dx, then we can understand that g is a function whose input is x and output is the slope of f at x.

Re: Mathematics for the Adventurous Self-Learner

#189
post #99

Earlier quoted context omitted.

The books are called "Essential Mathematics" 1 and 2. They correspond to these modules. http://www.openuniversity.edu/courses/modules/mst124 http://www.openuniversity.edu/courses/modules/mst125 Second-hand OU books are usually bought from https://www.universitybooksearch.co.uk/

Could you provide the direct links for the books used in: MST123, MST124, MST125? I could not find them in your links.

If you're not in the UK, I see that Thriftbooks appears to get at least some copies[1].

1: https://www.thriftbooks.com/browse/?b.search=MST124#b.s=most...

Re: Mathematics for the Adventurous Self-Learner

#190

As someone who dropped out of highscool after 10th grade and never went to university/college one great way I've found for learning mathematics without any foundational basis is trying to learn CG/3D programming. I always felt like maths was too abstract to keep me engaged, but when the output of your work is immediately observable visually it becomes a lot more engaging. There's just something so much more satisfyin…

This is why I love to do game development without using a game engine. It gives you a reason to learn math, optimize your code down to the metal, all while having fun playing your game.
Post reply on HN