aquafox

12 ms·

1. 121.▲
   by aquafox 4y ago

   You should learn R and dplyr ;)
2. 122.▲
   by aquafox 4y ago

   You do know that there are things like epigenetics, DNA repair (using specialized proteins), RNAi, post-translational modifications, metabolites (just to name a few)?
3. 123.▲
   by aquafox 4y ago

   > Technical founders often focus on a particular cool way of solving a problem and burn lots of capital building that thing. This nicely summarizes the current hype in Bitcoin or deep learning.
4. 124.▲
   by aquafox 4y ago

   Inmputing missing, in my opinion, is never a good choice. The main reason is that any subsequent method does not know which data were imputed and which not, i.e. it assumes the same error distribution. Using methods that can deal with missi
5. 125.▲
   by aquafox 4y ago

   Here is a much better explanation of PCA: https://stats.stackexchange.com/questions/2691/making-sense-... The key insight that many are missing is that PCA solves a series of optimization problems, namely that rec
6. 126.▲
   CRAN to remove 4747 R packages because of small issue in dependency package(twitter.com)

   1 points by aquafox 4y ago|0 comments
7. 127.▲
   by aquafox 4y ago

   As a cyclist, I can tell you that it's not about experiencing the challenge of climbing everest on a bike (and nobody who did it claims that). It is a challenge about gaining a very large number of vertical elevation on a single climb,
8. 128.▲
   by aquafox 4y ago

   The ECDF is particularly useful to compare two distributions. And it has the nice connection to the Kolmogoroff-Smirnov test for testing if two distributions are different: It's test statistic is the maximum distance between the two EC
9. 129.▲
   by aquafox 4y ago

   Stabile diffusion is most of all a waste of resources! 20 billion parameters trained, using I dont know how many CPU hours, all for making better artistic images than what an amateur would do and on par with professionals. Art is nice, but
10. 130.▲
    by aquafox 4y ago

    That's not true. There's a variety of reasons why outliers can exist, for example data quality issues where an experiment failed or the method to measure the data did not work. The approach presented there in my opinion has one bi
11. 131.▲
    by aquafox 4y ago

    I don't think a carbon tax will ever work in reality. Big companies are known for their tax evasions. We shouldn't bet the future of the planet on something that is known to not work.
12. 132.▲
    by aquafox 4y ago

    But the base 10 formula needs Bernoulli numbers, which themselves are defined recursively, meaning the required iterations are just hidden and not removed.
13. 133.▲
    The Cosmic Distance Ladder – Terence Tao(youtube.com)

    2 points by aquafox 4y ago|0 comments
14. 134.▲
    by aquafox 5y ago

    There's also an implementation in R: https://github.com/const-ae/einsum
15. 135.▲
    by aquafox 5y ago

    So many things wri g with it. PCA and SVD are placed on opposite nodes, but PCA is actually computed using SVD!
16. 136.▲
    by aquafox 5y ago

    Tried it, but Wifi speed on the Archer C7 was significantly reduced.
17. 137.▲
    by aquafox 5y ago

    I'm the one who made the original observation of the hidden network in the TP-link forum: https://community.tp-link.com/en/home/forum/topic/170160 Took a long time until TP-Link offered a firmware u