|
Rufo posted:CLRS is an obvious book suggestion for data structures and algorithms idk what the ~goon opinion~ of it is but its probably on like 99% of university reading lists while clrs is an excellent reference, dpv is very good imo while being concise and eminently readable the penultimate draft is also available fo free
|
# ? May 5, 2012 00:53 |
|
|
# ? May 9, 2024 23:18 |
|
clrs was the textbook for my algorithms class. Its probably the textbook for a lot of algorithms classes.
|
# ? May 5, 2012 00:55 |
|
Otto Skorzeny posted:while clrs is an excellent reference, dpv is very good imo while being concise and eminently readable nice. anal volcano u should read that ^
|
# ? May 5, 2012 00:57 |
|
clrs is also available for free if you know where to look. I rarely pay for textbooks.
|
# ? May 5, 2012 00:58 |
|
tef posted:I've failed more cs courses than you've had empty quotes no one likes anything i say so more than 0 isnt impressive. fail harder plz
|
# ? May 5, 2012 01:11 |
|
graphs are cool and graph theory comes up a lot in bioinformatics sure, i can assemble these genome reads into their shortest common superstring; just build a graph out of the strings with edges for their overlap, then find the heaviest hamiltonian ... poo poo
|
# ? May 5, 2012 02:30 |
|
Rufo posted:then this should definitely be your starting point. being able to understand formal logic and reading/writing proofs and WHAT A GRAPH IS is a fundamental skill in cs Kinda depends on where the calc I leaves off. Series come up often enough that being familiar with them is handy and sometimes that's like the tail end of a first-of-two-terms course, sometimes it's in the second-of-three-terms course in that sequence. I struggled some with math tbqh (got it well enough to finish a degree in the "hard" sciences but really had to slog through it), but when I went back to school more or less on a lark and took discrete math it was kind of a HOLY poo poo experience. It was the first time I ever really super enjoyed a math class and it seemed so useful day to day. For what it's worth, the course used Rosen's discrete math text, 6th ed. (The solutions manual is also pretty good, so it might be useful for self-study as well as in the context of a course.) Re: coursera being bootstrap, yeah, but they're like fresh off the starting block and seem to be spending most of their time working on actual course delivery stuff. Just finished the Algo I course with Roughgarden and it was pretty awesome. CLRS is kind of dense, but it was kind of the ultimate reference I'd turn to if something didn't make sense in lecture.
|
# ? May 5, 2012 02:32 |
|
Calc gets really important if yo'ure interested in machine learning algorithms: convex optimization is more or less "move in the direction of the gradient until you find your optimum" god help you if your optimizing something that isn't convex though; then you actually have to think about it SVMs with a gaussian kernel are heavy poo poo
|
# ? May 5, 2012 02:35 |
|
sign(sin(φx)) has one parameter φ but infinite VC dimension~ also I hate matlab
|
# ? May 5, 2012 02:38 |
|
Lysidas posted:Calc gets really important if yo'ure interested in machine learning algorithms: convex optimization is more or less "move in the direction of the gradient until you find your optimum" also doing a ML or MAP or whatever estimator for a lot of distributions will involve doing a fourier xform iirc
|
# ? May 5, 2012 02:45 |
|
when you get into category theory you find out that calculus concepts have their equivalents in the discrete realm that have applications but really who ever does that e.g. you can differentiate or integrate a data type and the result is useful
|
# ? May 5, 2012 02:45 |
|
Otto Skorzeny posted:also doing a ML or MAP or whatever estimator for a lot of distributions will involve doing a fourier xform iirc hey um this is kinda from the out field here but do you know anything about cycle/stability detection on periodic profiles. and if so where i can learn about more. i have likea 400 page book on linear control systems but i just want one specific subsection of all that (detecting when peaks and troughs are stable mostly)
|
# ? May 5, 2012 03:24 |
|
look at these fuckin nerdshoes just make an array and set up a couple for loops you're good
|
# ? May 5, 2012 04:01 |
|
Don't forget nyquist theorem hth
|
# ? May 5, 2012 04:06 |
|
While we're in algorithm talk, I'd like to ask if anyone has a few tips here. I'm working on doing some advertisement targeting at the moment and I've been working on implementing Interval IDs to evaluate Boolean Expression Trees (PDF) to allow me to do some nifty efficient arbitrary matches, getting close (~3x) to compiled-code speed when interpreted on a few reasonable queries. The problem I foresee is that it's still gonna be lovely expensive to run this stuff over hundreds of thousands of candidates. I was wondering if anyone working in different areas had similar problems of doing complex matching (hard to index) over medium-sized to large data-sets with ideally sublinear complexity, and if so, if you'd have any pointers . It seems to me pub/sub problems are somewhat similar, as well as match-making for dating sites or online games and whatnot. I'm sadly under low-latency constraints (under 20-30ms if possible), but any source of inspiration is welcome.
|
# ? May 5, 2012 05:12 |
|
MononcQc posted:While we're in algorithm talk, I'd like to ask if anyone has a few tips here. just use php
|
# ? May 5, 2012 05:23 |
|
graph algorithms and visualization are literally my favorite computer things
|
# ? May 5, 2012 06:31 |
|
MononcQc posted:While we're in algorithm talk, I'd like to ask if anyone has a few tips here. have you tried mongo db? It's web scale you know.
|
# ? May 5, 2012 06:34 |
|
MononcQc posted:While we're in algorithm talk, I'd like to ask if anyone has a few tips here. its weird, right? I never thought advertising would be the industry where I actually had to deal with serious cs poo poo but well here I am
|
# ? May 5, 2012 06:41 |
|
rotor posted:graph algorithms the 'send yostop' algorithm has a curious time complexity of O(5 months) that no one can seem to solve
|
# ? May 5, 2012 06:43 |
|
TiMBuS posted:the 'send yostop' algorithm has a curious time complexity of O(5 months) that no one can seem to solve this problem has been solved and it's integrated into the main trunk
|
# ? May 5, 2012 06:45 |
|
Jonny 290 posted:this problem has been solved and it's integrated into the main trunk why you gotta dis fork bomb like that
|
# ? May 5, 2012 06:52 |
|
rotor posted:its weird, right? I never thought advertising would be the industry where I actually had to deal with serious cs poo poo but well here I am I've been pleasantly surprised by how fun it is to work on that stuff. I used to think the projects I worked on were impressive because we had millions of requests a month and here I am handling more than that each hour.
|
# ? May 5, 2012 06:55 |
|
rotor posted:why you gotta dis fork bomb like that it was a yostop in trunk joke .
|
# ? May 5, 2012 07:01 |
|
MononcQc posted:I've been pleasantly surprised by how fun it is to work on that stuff. I used to think the projects I worked on were impressive because we had millions of requests a month and here I am handling more than that each hour. nobody likes a showoff
|
# ? May 5, 2012 07:06 |
|
i wish just a hundred people an hour would look at my company's website
|
# ? May 5, 2012 07:07 |
|
if it makes you feel any better between a third and a half of those requests are probably fraudulent
|
# ? May 5, 2012 07:10 |
|
yeah ok well last week ten guys from the philippines totally checked out our poo poo and they were impressed
|
# ? May 5, 2012 07:11 |
|
MononcQc posted:I've been pleasantly surprised by how fun it is to work on that stuff. I used to think the projects I worked on were impressive because we had millions of requests a month and here I am handling more than that each hour. Satisfaction of job well done Beers all around
|
# ? May 5, 2012 07:28 |
|
TiMBuS posted:hey um this is kinda from the out field here but do you know anything about cycle/stability detection on periodic profiles. and if so where i can learn about more. i was not the best student when i took signals & systems so i'm just going to handwave something to the effect of anything with a pole in the right hand side of the s-plane will be unstable. apologies that i couldn't be more helpful. i think mit has some EE opencourseware stuff that may be more useful than banging through a text yourself though.
|
# ? May 5, 2012 12:02 |
|
Rufo posted:what do you know about trees and graphs n poo poo? can u read a statement or proof written in propositional logic? how about using predicate logic to describe different properties of different types of graphs, e.g. acyclicity
|
# ? May 5, 2012 12:57 |
|
MononcQc posted:While we're in algorithm talk, I'd like to ask if anyone has a few tips here. im pretty stupid but that seems very similar to BLAST lookups, or at least, the underlying algorithm they use in BLAST may be worth checking out. there's way less variety in the search strings (ACTG only), but 10s or 100s of millions of characters of genome to search through
|
# ? May 5, 2012 13:46 |
|
nukethewhales posted:I can't imagine what the course you failed must have been like... probably one of those courses that requires you to write code instead of blogging about writing code
|
# ? May 5, 2012 14:37 |
|
Char posted:alternatively what do you know about software development i.e. the thing that real people do at real computers making real applications in the real world maths is hella relevant to non-trivial computing problems im sorry neat example: this poo poo is the reason its possible to do this poo poo on non-toy programs before the heat death of the universe (its cool. companies like nokia actually use it to analyse their code irl)
|
# ? May 5, 2012 15:39 |
|
math's
|
# ? May 5, 2012 15:42 |
|
Rufo posted:maths is hella relevant to non-trivial computing problems im sorry focusing on "computing problems" is fine if you're using your CPU as a gigantic calculator because it's 1952 and the only thing people write code for is crunching data
|
# ? May 5, 2012 15:49 |
|
in before "all programming is crunching data" hurf durf i sniff dijkstra's farts from a jar
|
# ? May 5, 2012 15:51 |
|
MononcQc posted:Interval IDs to evaluate Boolean Expression Trees lookit these fuckin shameful excel figures
|
# ? May 5, 2012 15:52 |
|
Char posted:in before "all programming is crunching data" hurf durf i sniff dijkstra's farts from a jar this except unironically. functional programming > *
|
# ? May 5, 2012 16:12 |
|
|
# ? May 9, 2024 23:18 |
|
also in lisp similar languages the program itself is even more data!!!! makes u think
|
# ? May 5, 2012 16:12 |