AppVeyorBot
commented
3 years ago AppVeyor build 1.0.249 for commit 0b590a39ef1cf53d5009e11bdd758e69322d185c is now complete. The rendered manuscript from this build is temporarily available for download at:Found 9 potential spelling error(s). Preview:
content/01.synopsis.md:6:TODO
content/02.intro.md:3:TODO
content/03.heterogeneity.md:51:TODO
content/05.prior-knowledge.md:14:TODO
content/05.prior-knowledge.md:21:TODO
content/05.prior-knowledge.md:35:TODO
content/06.multiple-approaches-required.md:14:autoencoders
content/06.multiple-approaches-required.md:14:TODO
content/06.multiple-approaches-required.md:14:VAEs...
allaway
jaybee84
jaclyn-taroni
Purpose
Addresses #111.
In an earlier pull request, I moved the coverage of DeepProfile and MultiPLIER into their own section (currently) titled "Multiple approaches are required." (Alternative title suggestions are most welcome!)
The goal of this section is to emphasize that you are most likely not going to employ one of the approaches we cover in isolation; you're going to need multiple techniques "woven together." A secondary goal of this section is to unify the regularization, other data and knowledge, and representation learning sections. Transfer learning is covered in the section immediately preceding this one and the main point of difference I'm trying to make is how putting some constraints on the representation the model learns can be helpful.
Directions for reviewers
I would like to probably take another pass through this before it goes in, but figured we could collect comments in parallel!
Which areas should receive a particularly close look?
Manuscript checklist
Unless otherwise noted above, this PR will be considered ready for review when all four items have been checked.