For those who use R and the tidyverse,
tidyr has been updated to
Some notable updates:
pivot_wider()provide improved tools for reshaping, superceding
gather(). The new functions are substantially more powerful, thanks to ideas from the
cdatapackages, and I’m confident that you’ll find them easier to use and remember than their predecessors.
hoist()provide new tools for rectangling, converting deeply nested lists into tidy data frames.
unnest()have been changed to match an emerging principle for the design of
...interfaces. Four new functions (
unchop()) reveal that nesting is the combination of two simpler steps.
expand_grid(), a variant of
base::expand.grid(). This is a useful function to know about, but also serves as a good reason to discuss the important role that vctrs plays behind the scenes. You shouldn’t ever have to learn about vctrs, but it brings improvements to consistency and performance.
So Google's gonna be Google. I was skimming through Google's AI blog (highly recommended btw) and was reading about some new neural networks, namely "weight agnostic neural networks" or WANNs https://ai.googleblog.com/2019/08/exploring-weight-agnostic-neural.html
So traditionally, you'll need to design the architecture of neural networks (i.e., how many layers, the connections, how many nodes, etc). These WANNs are apparently a way to use automation and have the computer find out which architectures work the best. It is a fascinating thing to think about.