jorgemf

52 ms·

481.
by jorgemf 9y ago
I like kotlin a lot, but suddenly there are a lot of articles in the front page about it. Kotlin isn't more cool because Google announced yesterday it is going to use it in Android. Kotlin is the same good language it was 2 days ago! (
482.
by jorgemf 9y ago
This is already in the language. It is just an interface which makes the common lists only visible as immutable objects. But inside they are a normal list. And in this case the object is read-only, not immutable, as we were discussing in th
483.
by jorgemf 9y ago
Sorry, I don't see the difference. With read-only you can create a new object which is the first element of the list and also points to the rest of the list. Instead of an structure you have an "immutable" object.
484.
by jorgemf 9y ago
I get your point now. Yes, it is read-only, not immutable. For me, I only need read-only. I think kotlin wasn't meant to be pure functional. I see it as object oriented with some functional programming. out of curiosity, what cases wil
485.
by jorgemf 9y ago
You can have immutability by making the setters methods of a class private. So if you combine it with "val" you have immutable objects. It is a bit more job in kotlin but as it is also object oriented programming you can achieve e
486.
by jorgemf 9y ago
> Kotlin has a `when` that desugars into if when statements Actually not. Not now, probably it was the case at the begining of the language. > algebraic data types they removed some restriction of sealed classes in 1.1. So I am not su
487.
by jorgemf 9y ago
> Scala needs to add a runtime when compiling to bytecode Last time I checked you needed like 500kb of libraries for any scala program compiled to Android. Kotlin is a few kb. Kotlin is a small addition to the java API, Scala needs much
488.
by jorgemf 9y ago
It is also because [] is an operator that you can override. I think because of this it would make more difficult to use it also to create arrays. It is more ugly for arrays but much more helpful for other things. And it also make sense with
489.
by jorgemf 9y ago
kotlin make it a bit simpler, once you do name != null you can use it as it is never null after than, there is not need to use let. It makes it more readable.
490.
by jorgemf 9y ago
pattern matching: swift has switf and kotlin has when, both are very similar recursive data structures: I don't get this one, kotlin has all the java api available and all the third party libraries. As java has been here for much longe
491.
by jorgemf 9y ago
If you have used kotlin, it seems the obvious move. I am not coming back to java, unless completely necessary. Kotlin is a great language (and more fun!).
492.
by jorgemf 9y ago
The real value of this is the algorithm of "one-shot imitation learning" (the paper is here: https://arxiv.org/pdf/1703.07326.pdf ). The title about robots it is only to catch the attention of the media. The
493.
by jorgemf 9y ago
> Snap generated $150M Yes, but Magic Leap generated 0 revenue. How can you be so sure it will generate something? It can be something like google glass, very cool but noone used them. Or the next smartwatch, some traction for a year and
494.
by jorgemf 9y ago
It can turn to be like google glasses, that it is very cool and promising but noone used them. So $4B valuation means 0
495.
by jorgemf 9y ago
$8B Valuation and no product in the market. It sounds crazy for me. Can somebody put some sense here for me?
496.
by jorgemf 9y ago
And for the job I don't know. You can post your CV here in the monthly thread for jobs, also take a look to the companies searching for people with experience in bioinformatics that wants to apply deep learning. I think a startup it is
497.
by jorgemf 9y ago
Cleaning the data, compare the models and understand why the results are like they are, those are huge things in Machine Learning. Actually it is like 90% of my job. Training the model it is nothing compared to it. As I said in another comm
498.
by jorgemf 9y ago
You are missing a lot of things that you don't know. If you want to do machine learning at some point you have to train a model. You need to know how to clean the data, how to create the train/validation/test set, how to meas
499.
by jorgemf 9y ago
I would say to you to go to kaggle and get your hands dirty in any competition related with bioinformatics. As you already know the domain better than other people in there you can focus on learning deep learning.
500.
by jorgemf 9y ago
I work in machine learning and I don't think a 3 months course is going to transform you in a machine learning engineer. I think you need at least the time you took. You need to learn the maths, statistics, etc. Apply those tools to re
501.
by jorgemf 9y ago
In academia almost anyone releases the source code of the experiments (at least it was this way 5 years ago, and things doesn't seem to have changed). If it is not true, probe me right. For example. NIPS proceedings (like hundred of pa
502.
by jorgemf 9y ago
> Please correct me if I'm wrong, smarter people, by replying. You are not an idiot, maybe not an expert but definitely not an idiot. Your description is quite easy to understand for someone without knowledge in the field. I would a
503.
by jorgemf 9y ago
There is a github link in the article with both things: https://github.com/facebookresearch/fairseq Google and Deepmind released a lot of stuff, I don't feel I have the right to complain about it.
504.
by jorgemf 9y ago
You can get a job with a portfolio in data science. Just go to kaggle and beat everybody in all competitions. That is worth more than a degree. Companies will try to reach you if you can do it. But, honestly, I think it is very difficult to
505.
by jorgemf 9y ago
> I don't think I could conceive of and write TensorFlow in 10 years of work Google was using for years a tool for deep belief networks, then suddenly deep learning was a thing and they started TensorFlow based on what they learn of
506.
by jorgemf 9y ago
You cannot do it before compiling the final executable, you have to do it while running. You can simulate the profiling but you will get the best results when you profile while executing. And I guess it is easier to do profiling in high lev
507.
by jorgemf 9y ago
> Questions: "Many optimisations that can really help high level managed languages like Kotlin simply don’t apply to C++": Which optimisations? I understood the JVM does an dynamic analysis of the code when running, so it can c
508.
by jorgemf 10y ago
Your comment make the Diogenes syndrome come to my mind. I prefer not to have something that I do not use it. I find it "distracting"
509.
by jorgemf 10y ago
That functionality also exits in some dumb phones. The experience is crap, but I guess it wouldn't matter if don't use it frequently.
510.
by jorgemf 10y ago
I think we have a problem with the apps getting bigger and doing more things. They use more resources and drain the battery faster. We have batteries bigger than ever, but still mobiles don't last for a day. I would want to go back whe

More ›