pilooch

33 ms·

331.
by pilooch 14y ago
Deep learning is so attractive to any AI and machine learning practitioner! The results are beautiful to witness or read about. This is clearly another step in a direction that many of us have been waiting (or working on) for a long time! T
332.
by pilooch 14y ago
I see you're a machine 'learnist'. If you feel like building your own 'intelligent' reader based on unsupervised learning, you can try https://xplr.com/developers/ (PM me for beta access). We're playing a bit with news as demo, see https
333.
by pilooch 14y ago
Just thought I would mention that you can use XPLR unsupervised machine learning API https://xplr.com/developers/ to build your own feed reader in the manner of https://xplr.com/products/illuminate/browser/ Though this probably applies
334.
Yandex shareholders to sell shares worth $600 million(reuters.com)
1 points by pilooch 14y ago|0 comments
335.
by pilooch 14y ago
There's this counter-intuitive trend that when the heritage of a dictatorial seat goes to a young one, everyone awaits for reforms, whereas the regime tends to rapidly increase its aggressive behavior. The same happened in Syria when Bachar
336.
Deploying security analytics as a service(techworld.com.au)
1 points by pilooch 14y ago|0 comments
337.
by pilooch 14y ago
Been discussing this with a friend today. Considering a short-sighted person, wearing prescription glasses: the problem is where to put the GGlass "screen". * it makes sense to put it inside, i.e. before the lens, because as short-sighted,
338.
by pilooch 14y ago
Question that comes immediately to my mind: I'm wearing glasses (short-sighted), so an augmented pair seems like an easy step, can I wear GGlass on top of my prescription glasses ?
339.
by pilooch 14y ago
> I don't understand why they keep pushing the niche GoPro-style use case Maybe because no one else sees you actually wearing the glasses :)
340.
by pilooch 14y ago
Sorry, I was actually surprised the URL went through without HN telling me it had been posted already. And not M$ ad intended here...
341.
Scroogle(d) by Microsoft(scroogled.com)
3 points by pilooch 14y ago|3 comments
342.
A subreddit recommender(xplr.com)
1 points by pilooch 14y ago|0 comments
343.
Ask HN: to give away (or not) your code after a Kaggle competition
2 points by pilooch 14y ago|0 comments
344.
by pilooch 14y ago
The last two paragraphs are hilarious, thanks for taking the time and writing this! Like anyone else I had my share of ridiculous shipment problems, but this guy remains playful (and cheerful!), wow :)
345.
A topic-based subreddit recommender (groups.google.com)
2 points by pilooch 14y ago|0 comments
346.
Ask HN: how does Google News selects its sources
1 points by pilooch 14y ago|2 comments
347.
Humanity doom in Computer Science buzz words(juban.free.fr)
1 points by pilooch 14y ago|0 comments
348.
by pilooch 14y ago
sure, I refer to computer science btw. Detrimental effects that I could witness: - Prof. are secured in their large offices, hard to get in touch with. This leads to useless work or idle students, waiting for feedback, or picking up the wro
349.
by pilooch 14y ago
The German hierarchy in academia is unbelievable. Few years back I moved from a federal lab in California to a state lab in Germany. I felt abashed at the extravagant praise of the head professor there. Also, the Prof. Dr. thing on business
350.
by pilooch 14y ago
You also can build a simple recommender based on categorization of articles. XPLR ( https://www.xplr.com ) for whom I work, has an API that among other things, allows to easily tag articles and build recommenders. If you're interested we'd
351.
by pilooch 14y ago
clickable http://http://deadsimplescreensharing.com/ Haven'tried. Suggested fix: on http://deadsimplescreensharing.com/#tutorial Comming Up -> Coming up
352.
by pilooch 14y ago
Oh boy, but are we missing the 'Golden Man' from Phillip K. Dick ? http://en.wikipedia.org/wiki/The_Golden_Man
353.
by pilooch 14y ago
LO Impress has been a pain recently. Saving a presentation and re-opening it would yield garbage. I've been testing LO 4 for a few weeks now, and it seems that the problem has gone. Apart from this, LO does the job, we use it for automatica
354.
by pilooch 14y ago
Other players, like blekko, may be more willing to open their API. Ask them directly.
355.
by pilooch 14y ago
Reading your comment, I thought I would mention this especially nasty bug in 4.7.1: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54075 Short description: a measured 3x performance regression from 4.6.2 to 4.7.1 when using std::unordered_ma
356.
by pilooch 14y ago
Actually sparsehash is not optimized for speed, but reduces the usual hash map memory footprint. In practice it is on par or slower than STL unordered_map.
357.
by pilooch 14y ago
There's more in the balance, Google is facing a huge tax case: http://www.france24.com/en/20121121-france-google-confident-... I am far from understanding the details of this, but it involves channeling money and contracts to Ireland, whe
358.
by pilooch 14y ago
Link to the paper: http://jmlr.csail.mit.edu/papers/v13/bergstra12a.html
359.
by pilooch 14y ago
good alternative to xmonad as it seems! bravo!
360.
by pilooch 14y ago
the one overview I've found the most useful is http://www.youtube.com/watch?v=ZmNOAtZIgIk (Bay Area Vision Meeting: Unsupervised Feature Learning and Deep Learning by Andrew Ng in April 2011).

More ›