osdf

From https://www.reddit.com/r/rust/comments/1pszdao/media_eilmeld...: eilmeldung is based on the awesome newsflash library and supports many RSS providers. It has vim-like key bindings, is configurable, comes with a powerful query language and bulk operations.

This proiect is not Al (vibe-)coded! And it is sad that I even have to say this.

Still, as a full disclosure, with this proiect I wanted to find out if and how LLMs can be used to learn a new programming language; rust in this case. Each line of code was written by myself; it contains all my beginner mistakes, warts and all. More on this at the bottom of the GitHub page.

There seem to be some nice developments into this direction recently? Something like the SfMNet (https://arxiv.org/abs/1704.07804) or the Correlation Filter Net (http://www.robots.ox.ac.uk/~luca/cfnet.html) for tracking.

There are several nice references within the article. I think the gvnn, Neural Network Library for Geometric Computer Vision, (https://arxiv.org/abs/1607.07405) should also be mentioned.

Possibly interesting: https://github.com/gregsexton/ob-ipython

If you like vim and spacemacs, give proton-mode a shot: https://atom.io/packages/proton-mode

Check my 'about' section here, should lead you to my mail :).

Thanks for your kind words! There is some stuff floating around, PM me.

No, sorry :( Clearly a TODO for some future work.

A lot of things happened since then. References up to spring 2014 are e.g. in our own work to super-resolve arbitrary sized images with convolutional en/decoders: Super-Resolution with Fast Approximate Convolutional Sparse Coding, http://brml.org/uploads/tx_sibibtex/281.pdf (which still has a lot of possibilities to be extended (e.g. color) and improved upon).

I had good results with http://www.willus.com/k2pdfopt/.

Thanks!!

Do you happen to have the project files (I guess pdf's and starter code?) available?

Or with some Machine Learning: Identifying Players in Broadcast Sports Videos using Conditional Random Fields, CVPR 2011.

Btw. Shane Battier (back then with the Rockets) was mentioned in a new york times piece on NBA and Stats some years ago: http://www.nytimes.com/2009/02/15/magazine/15Battier-t.html?...

Mentioned somewhere else: http://deeplearning.net/software/theano/ with a tutorial http://deeplearning.net/tutorial

Not C++ or Python, but lua with lots of stuff: torch 7 [[http://www.torch.ch/]]

Not to argue with you, robrenaud, but Hinton himself writes in their 2006 paper 'A Fast Learning Algorithm for Deep Belief Nets':

The greedy algorithm bears some resemblance to boosting in its repeated use of the same “weak” learner, but instead of reweighting each data vector to ensure that the next step learns something new, it re- represents it.

I guess that most people however would not think of this interpretation of greedily pretraining deep networks :). (I wonder if mbq had this in mind).

In the same article your point about good models of the input is mentioned, too (only copy&paste a small part of the paragraph):

Unsupervised methods, however, can use very large unlabeled data sets, and each case may be very high-dimensional, thus providing many bits of constraint on a generative model.

The 2006 paper is really an amazing read in my opinion.

There is a cleaner 'version' (==fork) of oh-my-zsh that fixed the long startup time for me: https://github.com/sorin-ionescu/prezto. With nice, fast setup (overwrites your .zshrc file!) and lots of documentation.

I guess it would have been nice if you just left your original text here and replied to my comment, pointing out how I misparsed a key point. Apart from that, (a) the Greece story is much more complex than how you try to frame it here (b) Greece and Spain are pretty different and shouldn't be thrown together. I also don't see a connection between communism and what has been discussed here under 'BIG' (I don't know of any communist country, actually -- mind you, 'communism' as it was/is meant!). And what 'socialist' countries grinding to halt do you think of?

As far as I understand most essays/reports from 'rather neutral' (yes, difficult to get an unbiased view) institutions here in Germany usually tell the story that nowadays a small minority lives on people's efforts, and this minority is well above welfare level.

Is there any 'proof' for your starting sentence? (countries don't last when ever more people rely on relative fewer producers). After all, automation levels increase ever more, and thus productivity, too. E.g. the number of people working in farming has shrunken dramatically (at least in first world countries), yet we have overall more than enough to eat ('overall'!). The fact that any country runs out of money is not an argument: Did Spain's productivity suddenly (or maybe also slowly) fell to zero? Did the people in Spain suddenly all lose their ability to work and think?

Sorry, tongue in cheek ... but do you say that you are self-sufficient? You are never using any public transport? Never drive on a street that wasn't built by yourself? Only eat things that you grew yourself. (No, it is not enough to counter with "I'm paying for it/I pay taxes.") First world societies are made up of people that _can't_ no longer be self sufficient.

So, you are saying that the majority (or, too many? how much is 'too many') of people in Greece and Spain are lazy? I don't understand 'pick up the tab', but the 'dwindling actual [german] workforce' will also become soon lazy? Why is the german workforce dwindling? Because of some 'lazyness' that's catching german workers?

And then these workers will cease doing unpaid work (which e.g. encompasses stuff like running theater groups, looking after the elderlies, ...)? Of course they'll do, because now they have to work on underpaid jobs to 'prove their worth' to society and get enough food on the table for their families.

It is widespread thinking here in Germany that those on welfare are 'keeping the welfare coming in and acquiring luxuries; they don't seem particularly motivated to work'. It is 'supported' by numerous shows (for whom? for those on welfare?) on TV.

It is on the other hand clearly refuted by a huge range of studies from various disciplines (economics, social studies).

Sorry, didn't see that the 'BIG' was mentioned in a different thread. http://en.wikipedia.org/wiki/Basic_income_guarantee

In particular for all free-market-lovers, 'fear of dying' makes the labour market (just as an example) pretty unfree ...

How a BIG may be financed can not be answered like this. You probably don't speak german, otherwise you may find this interview (and link to his doctoral thesis) interesting: http://www.forum-grundeinkommen.de/filme/andre-presse/doktor...

What it says is basically that for a BIG of 800 Euro (per month per person) in Germany there are only 20 billion Euro's missing (per year), that's nothing compared to e.g. what's currently spent for banks and other fancy things without blinking an eye (or in other words: our society is already working with such a high productivity that a BIG is already possible) The tax system (and social welfare system) needs to be changed completely, but the level of taxation will stay the same.

Could you be more specific with respect to 'empirical evidence'?

In Germany, according to the Federal Statistical Office (https://www.destatis.de), more hours are spent for unpaid work (100 billion hours) than for paid work (60 billion hours) (as of 2009). This gap seems to keep increasing.

Also in Germany, numerous (sorry, can't find a citation right now) polls showed that about 80% of the people still want (and will) work more or less fulltime with a BIG. Funny enough, also about 80% of the people think that with a BIG the _other_ people won't continue to do so and instead become (or are already) lazy.

With respect to LDA, there is a new paper on arxiv, Two SVDs Suffice: Spectral decompositions for probabilistic topic modeling and latent Dirichlet allocation [1].

[1] http://arxiv.org/abs/1204.6703

I'm curious about the hashing algorithm. I read about the planned whitepaper, but some preliminary info would be cool (e.g refs to academic papers that you build on).

Also, (basic) GPU operations with cudamat (or even finer, gnumpy) at your fingertips.

For a basic setup on OS X, I found scipy superpack to be an excellent choice: http://stronginference.com/scipy-superpack/

I like your comment. You probably are not German, but if, you would like the texts of Reinhard Kahl (http://reinhardkahl.de/). His texts are basically about the question 'What kind of school works', and actually 'What do we mean by "school works"'. Among others he tried hard to demystifiying the success of the finnish school system in the first PISA test, argueing against just copying blindly some aspects that _seemed_ important (e.g. like you, he pointed out that 'education' was socially highly respected in Finnland, unlike in most other european countries). He constantly writes about (german) schools that achieve very good results in tests like PISA, highlighting what they do differently. It is quite interesting that he often finds methods or techniques in these schools that fit really well a lot of findings in neuroscience and psychology (without the teachers at these school knowing anything about these findings!). So what are some things he advocates: (i) no longer teach in a fixed classroom setting, instead, every student should learn by himself, according to weekly/monthly/yearly plans that are setup together with his teacher(s) [main reason for this: every human is special, our brains do not develop at the same speed at the same time] (ii) do not kill the researcher in the children, instead foster playfulness in live and study and show that effort is something deeply satisfying (iii) Authentic teachers, authentic parents, authentic grown-ups (iv) Bring the 'real' live/world into the school (vi) open the school up to the 'real' live, schools should be centers in our society, (vii) at school you don't teach subjects, you teach human beings (viii) there is no school to rule them all and many more aspects. I like his stuff so much because he shows and describes places that really work, and also manages to write about philosophical aspects of these questions (he often connects to writings from Hannah Arendt about eduction).

PS: I'm aware that I can't give justice to the stuff Kahl writes, what I have written is quite incomplete. But one has to read it on its own. I'm also aware that some of this stuff is only relevant for german schools, yet I think the broader questions is relevant for all countries.

PPS: Refrain from calling me a hippy, a dreamer, a communist or any other stereotype _you_ think is negative. I went through the normal (german) school and university system and it gradually became clear to me, that things as they were and are working are not as good as we want to think, actually, things are rather bad. I'm looking for new ways and possible solutions, trying to find the right questions.