|
i just spent half an hour writing lovely database functions over std vectors in c++ and then another 20 minutes writing the first 7 questions out and debugging said lovely database functions. anecdotally, i noticed two things 1) not using an ide for c++ really does suck because of how much code you have to write 2) converting between a collection<T> to a collection<R> still sucks 3) maps iterator returning a pair<const Key, Value> can cause issues if you want to pull map's value_type's out into a vector. 4) template error messages are still awful. going to forget i did this in a few days and reattempt with clion. see if that helps.
|
#
?
Sep 15, 2014 06:01
|
|
|
|
| # ? Jun 5, 2024 15:57 |
|
|
i didnt do either of these things op
|
|
#
?
Sep 15, 2014 06:41
|
|
|
i too downloaded clion and have yet to use it
|
|
#
?
Sep 15, 2014 07:46
|
|
|
syntaxrigger posted:I've never read a book back to back on java. Last time I did work out of a java book was probably the java 6 deitel deitel book anyway in my mind there're two aspects to learnin a language/library/api/we. (1) learning what tools it gives you. (2) learning how those tools interact. so my usual approach is to sit down and read a big ole book on a language cover-to-cover, making notes on anything cool i come across. this helps w/ (1). then for (2) i write a small project of some sort in the language. this usually won't make use of even a fraction of what i learnt in (1), but it will correct many of the misunderstandings i've accumulated from, well, Learning Without Doing. if this approach appeals to you, then a good book for the bits n pieces of java is Java In A Nutshell. if you're completely new to the OO paradigm, you might want something heavier, but in that case someone else is gonna have to give you a recommendation. as for the project, welp, that one's entirely up to you.
|
|
#
?
Sep 15, 2014 08:00
|
|
|
re: (1): you can definitely get by using tutorial and stackoverflow questions, as half the loving industry demonstrates. problem is that without the context a proper survey of your toolset gives you, you frequently get hosed by the X->Y problem, which is that you encounter problem X, but don't know a neat tool exists for solving X. so you kludge together a hack that turns problem X into problem Y, which you can solve. and then somewhere down the line you realize your X->Y kludge is actually a piece of poo poo and you ask someone "uh how do i solve Y" and they go "why'd you wanna solve Y". when you answer "well i actually want to solve X, but i turned into Y", they laugh at you.
|
|
#
?
Sep 15, 2014 08:04
|
|
|
the issue with that is just the huge amount of possible known solutions to X for any language and one could not hope to grok it all from a single read of a single book. still better than going in with a "how to learn in 24 hours" tutorial and stackoverflow in 30 tabs tho
|
|
#
?
Sep 15, 2014 08:40
|
|
|
I don't want to say that the tasks were easy but I just wanna point out that ML in general is waaaaaaaaaaay harder than just that.
|
|
#
?
Sep 15, 2014 08:56
|
|
|
MeruFM posted:the issue with that is just the huge amount of possible known solutions to X for any language and one could not hope to grok it all from a single read of a single book.
|
|
#
?
Sep 15, 2014 09:00
|
|
|
what's ML? i want to say machine language but that doesn't seem to fit in this context. also im an idiot i've failed java in uni twice now, half the reason being i didnt go to class and the other half being me loving retarded at moving data being classes. so yeah, dont be so hard on yourself.
|
|
#
?
Sep 15, 2014 09:02
|
|
|
Symbolic Butt posted:I don't want to say that the tasks were easy but I just wanna point out that ML in general is waaaaaaaaaaay harder than just that. basically like stats has always been saucepanman posted:what's ML? i want to say machine language but that doesn't seem to fit in this context. also im an idiot
|
|
#
?
Sep 15, 2014 09:06
|
|
|
ML is often used as a buzz word for anyone manipulating data that can't be stored in an excel file without crashing. See also: big data. What I remember of ML is beysian statistics, clusters, markovs, vector machines, etc. So much mathy bullshit (not really bullshit, just hard for people without a good math foundation like me.) I have yet to see any person outside of academia with a data scientist-like title use these techniques.
|
|
#
?
Sep 15, 2014 09:09
|
|
|
Most of the time, they're just programmers with an advanced degree and tableau experience.
|
|
#
?
Sep 15, 2014 09:11
|
|
|
coffeetable posted:ehh. there are a much larger number of people who'll happily bang a SVM/k-means/whatever implementation against problems than the number of people who actually have a good grounding in the field. but the thing is that there is a much, much, much larger number of people who haven't touched a ML and are scared ever to, so it's not v hard to look like a 'ML person' in front of them cool, what you're saying is adding to my confidence in the case I ever find a job opening involving ML. I'm totally an "ML person" because I worked through some courseras  btw syntaxrigger additionaly to reading a book you should look up those data science, machine learning online courses and try to work through them with java
|
|
#
?
Sep 15, 2014 09:19
|
|
|
haha all this time i thought ML was referring to OCaml, SML, etc. learn something everyday 
|
|
#
?
Sep 15, 2014 11:17
|
|
|
saucepanman posted:what's ML? i want to say machine language but that doesn't seem to fit in this context. also im an idiot go 2 class
|
|
#
?
Sep 15, 2014 12:34
|
|
|
Shinku ABOOKEN posted:haha all this time i thought ML was referring to OCaml, SML, etc. Same, I was about to knock up a haskell solution to those problems to demonstrate how difficult ocaml would be from java
|
|
#
?
Sep 15, 2014 12:35
|
|
|
thanks for the python posting tef
|
|
#
?
Sep 15, 2014 12:49
|
|
|
BONGHITZ posted:go 2 class yeah this
|
|
#
?
Sep 15, 2014 14:21
|
|
|
fart simpson posted:actually that was a lie. i learned what monoids and functors are, because those are both really simple. then i read that monads are just monoids in the category of endofunctors and it was obvious
|
|
#
?
Sep 15, 2014 14:50
|
|
|
Symbolic Butt posted:btw syntaxrigger additionaly to reading a book you should look up those data science, machine learning online courses and try to work through them with java java has the ultimate library for this: apache mahout mahout is not a focused or carefully designed library. it's just every drat ML algorithm in one giant grab bag. sometimes that's ok
|
|
#
?
Sep 15, 2014 15:13
|
|
|
Notorious b.s.d. posted:java has the ultimate library for this: apache mahout see also: scalaz
|
|
#
?
Sep 15, 2014 15:37
|
|
|
wait gently caress we're talking about Machine Learning still
|
|
#
?
Sep 15, 2014 15:37
|
|
|
Looks like I might get to drop a class I had chosen to take next year, which means either goodbye functional programming class or goodbye...advanced networking. Im not 100% sure what I would I would get rid of
|
|
#
?
Sep 15, 2014 15:42
|
|
|
i feel bad about piling on syntaxrigger but data prep is critical for any kind of ml and that assignment looked exactly like the sort of things you do on first passes thru the data to get a feel for it but dang i wouldnt do it in java
|
|
#
?
Sep 15, 2014 15:45
|
|
|
networking is a nightmare functional programming might be a little better?
|
|
#
?
Sep 15, 2014 15:47
|
|
|
Symbolic Butt posted:cool, what you're saying is adding to my confidence in the case I ever find a job opening involving ML. I'm totally an "ML person" because I worked through some courseras drat it I keep forgetting about coursera, I want to finish a few of these before I graduate and have to get areal job Vincent K. McMahon posted:Looks like I might get to drop a class I had chosen to take next year, which means either goodbye functional programming class or goodbye...advanced networking. Im not 100% sure what I would I would get rid of I would take networking because I assume it also covers some hardware topics which is always interesting
|
|
#
?
Sep 15, 2014 16:16
|
|
|
I would do networking disclaimer: i am retarded and actually enjoy it
|
|
#
?
Sep 15, 2014 16:46
|
|
|
yeah i did a cisco unit and it was pretty drat fun, but you have to like networking
|
|
#
?
Sep 15, 2014 18:42
|
|
|
Sorry yeah, networking last year was ok. Got lucky on the exam though. OK so the classes I have just now are..one about teaching CS in highschool, where I spend 10 half days in a classroom being a teaching assitant, and think I might need to teach 1 lesson. No exam for this one. The others are a class on computer architecture, which looks like hardware stuff obviously, seems fun. Adv Operating sytems, which is a class ive been told is good and again has a lot of hardware aspects to it. A class for sofware engineers workin in the financial world, taught by my future emloyees. Then there is adv networking. Cyber security (not sure on this one), functional programming (not sure on this either). oh yeah and an AI class, which seems fun but also really hard
|
|
#
?
Sep 15, 2014 19:02
|
|
|
Vincent K. McMahon posted:Sorry yeah, networking last year was ok. Got lucky on the exam though.
|
|
#
?
Sep 15, 2014 19:14
|
|
|
also pick whichever course has the most maths in it
|
|
#
?
Sep 15, 2014 19:14
|
|
|
Vincent K. McMahon posted:Sorry yeah, networking last year was ok. Got lucky on the exam though. have you done fp before? do you know what language the class is taught in?
|
|
#
?
Sep 15, 2014 19:18
|
|
|
coffeetable posted:also pick whichever course has the most maths in it do this in the real world no one will give you time or motivation to do maths. so get it done in uni so that it stays in the back of your head when you do stuff in the future.
|
|
#
?
Sep 15, 2014 19:19
|
|
|
coffeetable posted:link your AI syllabus By the end of the course students will be able to: 1: Demonstrate familiarity with the history of AI, philosophical debates, and understand the potential and limitations of the subject in its current form; 2: Map the basic components of an intelligent agent, and be able to map these onto other advanced subjects such as information retrieval, database systems, computer vision, robotics, human-computer interaction, reactive systems etc; 3: Implement basic search-based problem solving techniques; 4: Describe basic issues in planning; 5: Describe basic utility theory and a probabilistic framework for rational decision making; 6: Create and use simple belief networks using available software packages; 7: Straightforward machine learning techniques to learn from data; 8: Discuss the difficulty of computer perception. Ploft-shell crab posted:have you done fp before? do you know what language the class is taught in? I done a raelly horrible group project last year using Erlang, but that is the extent of my fp knowledge. First week or two of the course is looking at something called livescript and then remainder is using haskell Shinku ABOOKEN posted:do this remember im looking to drop one of these classes. I done enough math classes in 2nd year, barely passing some of them, to realise its probably not for me
|
|
#
?
Sep 15, 2014 19:24
|
|
|
Vincent K. McMahon posted:Cyber security (not sure on this one), functional programming (not sure on this either).
|
|
#
?
Sep 15, 2014 19:28
|
|
|
Soricidus posted:both of these could be interesting if they're done well, or hellish if they're done badly. cyber security is more likely to be useful in the real world, if that's a consideration. I know the coursework for cyber security is worth 20% and entails being given some encoded text and you need to work out the cipher for it lol but yeah it seems like it would be useful. I am going for a mix between easy classes and things that are useful and interesting
|
|
#
?
Sep 15, 2014 19:31
|
|
|
Given the choice, I think unless I had a prior interest in networking, I would take the fp class, because it will probably make you a better programmer. e: if the choice is between fp and advanced networking
|
|
#
?
Sep 15, 2014 19:32
|
|
|
why're you intimidated by that ML syllabusVincent K. McMahon posted:I done enough math classes in 2nd year, barely passing some of them, to realise its probably not for me
|
|
#
?
Sep 15, 2014 19:32
|
|
|
coffeetable posted:why're you intimidated by that ML syllabus for the amount of effort it looks like ill need to put in, im not sure if i am THAT interested or if i will get a few weeks in and then just kind of go idle we actually do have a proper ML course too, By the end of the course students will be able to: 1:Discuss the principle of learning from data; 2:Describe the main machine learning methods: regression, classification, clustering, probability density estimation and dimensionality reduction; 3:Use a selection of common machine learning algorithms and be aware of when one is to be favoured over other; 4:Implement and use machine learning algorithms in Matlab; 5:Discuss the major machine learning application area in, for example, Information Retrieval, Human Computer Interaction, Bioinformatics and Computer Visions & Graphics; 6:Detail emerging machine learning approaches such as non-parametric methods and sampling techniques.
|
|
#
?
Sep 15, 2014 19:42
|
|
|
|
| # ? Jun 5, 2024 15:57 |
|
|
syntaxrigger posted:So an ML shop contacted me because I have java in my resume and I have lofty goals of one day not sucking at ML. They sent me a programming test that I did horrible in because I am not strong enough in when to use what data structures and algorithims. So instead of wallowing coffetable convinced me to post here for some insight into what I should have done. if the company is well-known or has room to be picky with candidates, chances are they don't really just want a "sum these up, sort and return the first 10" solution. they are probably looking for candidates that know to use a heap to do the sorting in n log k, and probably someone that does all the counting in parallel.
|
|
#
?
Sep 15, 2014 21:48
|
|