Live data from Hacker News

CS 168: The Modern Algorithmic Toolbox

web.stanford.edu

81–90 of 101 posts

Re: CS 168: The Modern Algorithmic Toolbox

#81

VMLS by Stephen Boyd (also of Stanford) would be great background reading for the vector/matrix stuff (i.e. most of this course): https://news.ycombinator.com/item?id=18678314 I love this book. As opposed to the way that linear algebra is typically introduced, this book focuses on concrete applications (like text analysis, image/signal processing, finance, ML, etc.) and eschews more arcane concepts (like eigen ). To…

Another excellent paper on the similarities between all those linear models is “ A Unifying Review of Linear Gaussian Models”

> Factor analysis, principal component analysis, mixtures of gaussian clusters, vector quantization, Kalman filter models, and hidden Markov models can all be unified as variations of unsupervised learning under a single basic generative model.

https://authors.library.caltech.edu/13697/1/ROWnc99.pdf

Re: CS 168: The Modern Algorithmic Toolbox

#82

How can you cover such a huge and diverse amount of material with any level of rigor? I view this type of courses as pointless. There's not enough time to cover any problems or pitfalls.

Do you have any one particular topic in mind you think isn't covered rigorously enough? If so, what's missing from it? That way someone can use "is that covered" as a heuristic for evaluating other resources in the future.

One week (two lectures, one discussion, one homework) for Fourier transforms?

Seriously?

Having been a TA for a course with similar pretentions. Students will be miserable because they aren't given the tools to deal with the homework, they will learn very little, and forget it all very quickly

Re: CS 168: The Modern Algorithmic Toolbox

#83

Earlier quoted context omitted.

Do you have any one particular topic in mind you think isn't covered rigorously enough? If so, what's missing from it? That way someone can use "is that covered" as a heuristic for evaluating other resources in the future.

One week (two lectures, one discussion, one homework) for Fourier transforms? Seriously? Having been a TA for a course with similar pretentions. Students will be miserable because they aren't given the tools to deal with the homework, they will learn very little, and forget it all very quickly

Stanford has a course dedicated to FFT, which is available online [1].

[1] https://see.stanford.edu/course/ee261

Re: CS 168: The Modern Algorithmic Toolbox

#84

Crickey, what an exciting syllabus. What I would give to be a full-time student again and study stuff like this without any distractions.

Even when I was a student and loved a particular class, there was the constant distraction of having to attend and do homework for four or five other classes.

Still much better than the current situation of 8 hour day job and family.

Re: CS 168: The Modern Algorithmic Toolbox

#85
post #65
post #38

Earlier quoted context omitted.

try a side project? when I think of fun ideas, I always notice how easy it is to bump into computationally hard or complicated stats/ML/math problems left and right, even if it's just a simple game. I remember one time when I realized this: a TV show character suggested building an app that lets you take a picture of someone's shoes and then give you links where you could buy them. It sounded a) like a pretty good bu…

I suppose even then you'd use a library which solves the problem rather than figure out and build the entire thing yourself. I do agree that algorithms are fun and the basic ones are probably useful in everyday work, but for anything complicated you'd rather use a proper library. It's also definitely good to have knowledge of their internal workings though.

with anything involving AI/ML, like the shoes-app, you'll likely still have to create or train a model to achieve exactly what you want it to do.

Re: CS 168: The Modern Algorithmic Toolbox

#86
post #38
post #33

It would be cool to be an expert at algorithms, apply them to hard problems, and build cool things. I feel like, for me, there's no point in learning them since the work I get is always "put a button here, update the DB" type stuff.

try a side project? when I think of fun ideas, I always notice how easy it is to bump into computationally hard or complicated stats/ML/math problems left and right, even if it's just a simple game. I remember one time when I realized this: a TV show character suggested building an app that lets you take a picture of someone's shoes and then give you links where you could buy them. It sounded a) like a pretty good bu…

Is this now a trivial problem because of Deep Learning?

Re: CS 168: The Modern Algorithmic Toolbox

#87

Earlier quoted context omitted.

Here’s one explanation of Consistent Hashing by a popular author that just started a systems design YouTube channel: https://m.youtube.com/watch?v=UF9Iqmg94tk

I love this channel, I've been watching their videos for a few weeks. The guy has a great ability to break things down and explain complex topics with ease.

Haven’t looked at his books yet, but they appear to be very popular:

https://www.amazon.com/Alex-Xu/e/B08BNMFT7P/

He also has a cheap online course ($60 a year) that covers everything that’s in the books:

https://bytebytego.com/pricing

* note: not affiliated with him, YouTube suggested his videos and I looked into him.

Re: CS 168: The Modern Algorithmic Toolbox

#88
post #30

Earlier quoted context omitted.

> I was thinking that the vast majority of those topics should be pretty much standard knowledge for a mathematically trained computer scientist Well yes, this is literally that mathematical training.

yeah, I guess I was expecting postgrad-level stuff on things that are curious to people who finished a CS education

I doubt even a bare majority of HN readers have finished a CS education.

Re: CS 168: The Modern Algorithmic Toolbox

#89
post #15

Earlier quoted context omitted.

...or, you could just 1) use the raw count of inbound links, weighted by the count of inbound links of the linking pages, normalize by the total number of links on that page, and run it a few iterations to get a first approximation of the probability distribution. This is a pretty intuitive approach, would almost certainly have been good enough for Google, and avoids all the jargon, probability, and "well studied alg…

0. Why so anti-intellectual? 1. The algorithm you're attempting to describe in 1) above is either exactly PageRank or an approximation to it, so I don't get the point you're trying to make. 2. The point of jargon and well studied algorithms is that you can recognize when your problem is a problem someone else has already tackled. In the specific case of PageRank, if you recognise it's an eigenvalue problem you can 1)…

0. Hah, yeah. Pretty sure you’re the first person to accuse me of being anti-intellectual...I can’t say for certain, but I think this interpretation here might be on you.

1. Indeed I think you missed my point: the causality implied in your earlier post reads as “well, if you know math, PageRank pops right out!” — my point is, you don’t need to know anything about eigenvalues for PageRank to be an intuitive solution; the insight isn’t eigenvalues or even probability distributions of pages, it’s ranking by relative rank of inbound links. There are many ways to operationalize that insight, and your post made gatekeeping allusions about what math it takes to do that. That’s what I’m responding to.

2. Sure, but you were originally responding to a post about when graph algorithms might ever be useful in industry. Obviously all things being equal, knowing about graph algorithms > not knowing about them. But again, you don’t need to know about them in advance.

3. No doubt Google’s early success is due to PageRank. But PageRank’s success is due to the insight behind it — following web links backward (in fact the precursor to PageRank was called “backrub”) — and not due to eigenvalues.

To be clear, I’m not at all anti-intellectual, and I love a good math puzzle or algorithm as much as the next guy — and I said as much in my other post.

But I’m also not in love with the fetishization of formalisms that is common in technical and academic circles, and the attendant value system that is so eager to prove itself that it tries to fit any and all adjacent insights into its own paradigm — just as you’ve done here.

Re: CS 168: The Modern Algorithmic Toolbox

#90
post #15

Earlier quoted context omitted.

...or, you could just 1) use the raw count of inbound links, weighted by the count of inbound links of the linking pages, normalize by the total number of links on that page, and run it a few iterations to get a first approximation of the probability distribution. This is a pretty intuitive approach, would almost certainly have been good enough for Google, and avoids all the jargon, probability, and "well studied alg…

I'm not into mathematics or algorithms and the like at all, but if someone explains things like PageRank in layman's terms, I'm like, "yeah I get that". That's been the recurring theme with me and mathematics, if it's just the theory, language and formulas I'm like "what?", but give me a practical example / use case, a way to visualize it, and I get it. I struggled with linear algebra in school - I don't even know wh…

You are not alone.

In my experience, there’s a language of mathematics that makes an almost comically strong attempt to make itself inaccessible through jargon, single-letter variable names, and an over reliance on symbolic manipulation.

If you talk to most mathematicians or algorithm researchers, you’ll find that they are some of the most intuitive people you’ll ever meet, and they use visualization and other intuitive techniques all the time.

And then the classes are all about symbol manipulation.

I swear there are millions of people out there who could be mathematicians but just couldn’t get excited by all the symbol manipulation.

Post reply on HN