c0g

I don't think so, Slovakia isn't on the skills list[0]

[0] http://travel.state.gov/content/visas/english/study-exchange...

Having just got a J1 visa myself, he'd be well served with J1. It lets you work in the US for up to 6 months in an internship. Might still be time!

coff MCMC coff

> motorcycle

I was tangentially involved with the Kew Gardens mosquito project - however they're all exciting and worth a read!

Hi Viral, do you have an intuition for why Numba would be faster than Julia here? It appears Julia spends a lot of time calculating the log on line 71.

I love how you're getting downvoted when the truth is NNs work well _only_ when you have a team of specialists to select their hyperparameters. There's a rich vein of research for AutoML to automatically learn the hyps.

If you live near enough to an exchange for ADSL over BTWholesale's elderly copper wires to be fast or in a big enough town for Virgin/BT to have laid cable/fibre, your net service is pretty good in the UK - from what I hear better than in the US.

If you live further away so Wholesale has no competition from Virgin they don't lay fibre or maintain the last mile and non-town service is (again, purely anecdotally here) worse. Non-town is not rural, it's maybe 3 miles away from an exchange.

London ain't the UK, basically.

Machine Learning is more of an engineering discipline than a philosophical one. The aim (mostly, with exceptions) is to make machines that learn, not to research the nature of "understanding" or "concepts".

There are people who are looking at broader ideas of "what is intelligence", sometimes even the same people doing the more statistical research, but it's an open ended problem.

The main reason science/math/statistics doesn't spend time worrying about philosophy is because you can't really specify the problem cleanly - and having a clear idea of the problem is 90% of the way to finding the solution.

[edit]: If you want brain-inspired ML stuff, it's worth looking at what neuroscientists and cognitive scientists are doing. From what I've seen of some of their work they're developing ML algos as minimal tests for understanding how parts of the brain or consciousness works.

Poorly - Julia calls well tested and really carefully programmed BLAS libraries for the numerical heavy lifting, this doesn't. It is likely to be less accurate, less stable, and slower just for that reason.

The comparison doesn't really work anyway - better would be Julia vs. Go, some ML toolkit in Julia vs. this.

Allowing for food, transport, etc?

Yes! Numpy is hugely popular in (at least) physics and machine learning and the main negative comment I get is "Wow that's horrible syntax". I'm still not sure they'll be convinced, however:

    trans_coef.T.dot(solve(H.dot(V).dot(H.T)), trans_coef)

becomes

    trans_coef.T @ solve(H @ V @ H.T, trans_coef)

but in MATLAB or Julia might be:

    trans_coef' * ((H * V * H') \ trans_coef)

....I want a backslash operator now!

Given you're Romanian, that means you can easily study in other European countries right? I know of great places in Germany, France, the UK and Finland to study - and there are definitely more.

Might be worth considering - get somewhere with a big tech scene, higher wages and more difficult courses.

Yeah, I normally do two swipes then rinse to get rid of my manly beard deposits.

It's a safety razor?

http://en.wikipedia.org/wiki/Safety_razor

Thanks for the pointer, I shall play with latexml!

Specifically for writing math heavy papers (what I use LaTeX for) HTML offers:

1. (kinda.. if I want to use Google's font API and the browser renders it well)

2. Not a single file.

3. Not cross-platform - it will render differently on my linux desktop and mac laptop.

4. I don't know how to convert all of my LaTeX macros and workflows to HTML so I'd need something like htmltex (doesn't exist, I want it).

5. MathJax is good, but it ain't great. Too many compromises.

6. No typesetting and words per lining heuristics like LaTeX - again, need htmltex.

7. and 8. were a bit unclear, sorry. I mean in writing mathematical papers. HTML has a huge community but signal to noise for writing readable papers is pretty high. Mapping what I do in LaTeX to HTML requires many more lines of code.

What would be great is if LaTeX compilers could output to HTML or other more portable backends. I've grown to love LaTeX since my undergrad- I don't have to worry about how things will display like I do with markdown, html and MathJax, it just works. I don't know how possible that will be to replicate in a browser.

PDFs stink, but I've yet to find something that stinks less!

What I need:

* High quality fonts and images

* Single file

* Cross platform

* Simple to create using LaTeX

* Excellent support for math

* LaTeX style readability considerations

* Massive user base

* Many extensions

Can I call a Java library (say, google play services) from Qt without having to jump through JNI hoops?

As a pedal- and motor- cyclist in Oxford (city big into bikes in the UK), these electric bikes really get to me. We have lots of amenities for pedal cycles to improve their experience of the roads which are (rightly) unavailable to the faster and more dangerous to pedestrians motorcycles. Seeing these electric assistance bikes caning it along at an unrealistic speed by disinterested and distracted riders makes me think they should be classed as motorcycles with all the restrictions/responsibilities.

Edit: For example, riders of a motorcycle or moped must carry insurance and wear a helmet. Anything over 50cc and you need to pass a training course to teach the rules of the road and basic manoeuvres. Cyclists do not need to pass anything (or pay vehicle tax...). I'm unsure about my opinion on making human powered vehicles abide by licensing/insurance requirements, but powered vehicles really should do.

For backporting to 3.3/2.7. 3.4 has it.

I think the interesting comparison will be between something like Tegra 5/Exynos 5 and this. Most scientific computing these days is CUDA, and getting something like cuBLAS onto this chip will be a struggle. The Exynos/Android combo has already lowered it's ability to interest me by not even accepting OpenCL as a first class citizen.

EDIT: Hopefully NVidia will give me CUDAndroid, so I have fun.

    Nonetheless, Epiphany does see a few situations where it     might perform better than a classic CPU and better than a classic GPU. Its just a different software architecture, focusing on a different problem niche.

Do you have any examples?

Make it interface with Python and Matlab on my local machine and export to .ps and I'll give you some money.

He already said he thought they were morally required to release the code on the net. Also, what are you talking about?

Yes it was :) but time well spent imo.

On reflection I guess I might have had more free time to spend on this than a normal person - I did the SVM as a [small] part of my masters project, so if you're time constrained with a real job and a life then might be best to disregard me.

Speaking as a PhD student in machine learning-

Implement the algorithm yourself, first, in Python+Numpy. The only reason I feel comfortable with Gaussian Processes and SVMs is due to writing code to solve them manually.

Once you're happy with the basics, and can test your ideas with code you intimately understand, optimise for speed by using a library like this.

Computer science and software engineering are different things. Computer science is a specialised subset of mathematics, software engineering is the process of making things that run on computers. They're connected in the same way that civil engineering and physics are- one underpins the other, and maybe in time what the physicists are doing will help make a good bridge. Science underpins engineering, engineering justifies science.

I have had no trouble compiling from source. Are you trying to use a PPA?