Publication in Significance – code

Couple of months ago I published a paper in Significance – couple of pages describing the essence of deep learning algorithms, and why they are so popular. I got a few requests for the code which generated the figures in that paper. This weekend I reviewed my code and was content to see that I used a pseudorandom numbers, with a seed (as oppose to completely random numbers; without a seed). So now the figures are exactly reproducible. The actual code to produce the figures, and the figures themselves (e.g. for teaching purposes) are provided below.

Deep learning algorithms

Reference: What’s the big idea? Deep learning algorithms.

3 comments on “Publication in Significance – code”

  1. Thanks for sharing this! As a note to other readers who might get misled by the usage of the term “pseudorandom” in this article: It’s not the absence of manually setting a seed that defines whether numbers are truly random or pseudorandom but whether the numbers are computationally deterministic. Base R always produces pseudorandom numbers – if one doesn’t manually set a seed, R will automatically generate one. See ?base::Random: “Initially, there is no seed; a new one is created from the current time and the process ID when one is required.” Setting a seed manually makes the pseudorandom numbers *reproducible*, which is great. True random numbers can be obtained e.g. by using the package {random} which obtains them from RANDOM.ORG and thus from atmospheric noise.

Leave a Reply

Your email address will not be published. Required fields are marked *