Select count(*) takes a surprisingly long time, even on an indexed table. Apparently thus is because of the MVCC model and row count estimates are available to mitigate.
My understanding is that this problem was alleviated with index-only scans in Postgres 9.2. Using these, each tuple no longer needs to be visited to check for visibility.
Index only scans work in some cases but not all. In order for an index only scan to be more efficient than a regular index scan, the visibility map needs to be up to date. This is a problem because (currently) the only thing that updates the visibility map is the vacuum and the autovacuum only runs after a large number of updates/deletes happen on a table. If your table is largely read/insert only, the only way to keep the visibility map up to date is to manually call VACUUM on your tables periodically which is a huge pain.
Are the newer ships more environmentally friendly? I'd assume only slightly because they're still burning bunker oil, only slightly more efficiently. Which I find perverse given the attention focused on car pollution.
Much more, yes - modern ships have been built slower, with lover rev-engines, to save fuel. And a larger ship is inherently more environmentally friendly. And even the oldest ships do far, far less environmental damage per kilogram per kilometer than cars or other road transport. Not to mention that the issues with nitrogen oxide emissions in particular (the focus of the recent scandal) are less about environmental damage and more about the danger to humans from breathing them in, which is obviously a lot more important when emissions are happening in the middle of a city rather than out on the open sea.
Congratulations on hitting 1.0, Keras really looks like the way forward for python ML. I tried to use Keras after doing Andrew Ng's Coursera course but found that I just couldn't connect the dots from what I'd learned on the course to the concepts assumed familiar by the keras documentation though.
Yes, I think we should work on a FAQ to introduce common ML concepts and their implementation in Keras. Any specific concept that you had trouble with?
I came across that glossary a few days ago when looking to solve an RNN problem with a deadline. You may be pleased to know I think I may have solved it by switching to Keras, and this post actually helped unsticking me from a smaller problem too.
Thanks, and congratulations on version 1.0!
Cool. It sounds dumb but just shaping the inputs was a problem for me. I wanted an LSTM/RNN but epochs, batches and recurring sequences didn't click together for me. I also had trouble plotting the learning curves.
All the above things are possible only when the other parts of the interacting system - The Insurance System/Phone Contracting Company begin building relatively user/AI friendly interfaces.
Majority of these (mundane yet important)tasks go through poorly designed systems and expecting AI to navigate that hell-hole is a far fetched dream given the current climate (Maybe, a dozen years). On the other hand, if we slowly move towards a world where these systems are also replaced by AI, then we will be closer.
I was just recently wondering if an LSTM neural net could be trained as a decent decompiler from LLVM IR to Rust source. Feed that with Clang output of an established OS and coreutils(i was thinking minix3 but maybe seL4) and get a working if not necessarily idiomatic Rust OS. Refactor at leisure.
Like other commenter said, it would be unreadable. You'd be better off using a transformation system like Semantic Designs DMS or Stratego language with heuristics. Series of rules, analyses, and transforms applying until new code exists.
Real problem with your scheme is that the typing and structure are important to effectively using Rust. Those that you get in LLVM will be nothing like what you would want in Rust and you'll need additional information. In other words, source to source with annotations is a better option.
That's because the Elm compiler doesn't yet do much in the way of "dead code elimination" (DCE).
However, that will change in a future release of Elm, once Joey Eremondi's work has been fully integrated. My understanding is that integration is not slated for the 0.16 release (imminent) but will likely be part of the 0.17 release.
Are they planning to leverage the Google Closure compiler the way Clojurescript does? Un-optimized Clojurescript is also huge until it runs through Closure compilation.
No, I believe the DCE implementation (still in the works) is specific to the Elm compiler, as opposed to an implementation which organizes the source in such a way as to leverage Google's Closure compiler.
I'm genuinely interested to know if the altruistic/community driven ethos of these efforts ever last? Once the NGO leaves town and the local federalè or whoever notices a nice little income stream?
'At this price, the community reaps a profit of about 49,000 pesos, or $3,600, per year. The community has appointed a committee to manage the incoming funds, setting aside some money for maintenance and repair of the system, and investing the rest back into the community.'
