In #114, I proposed restructuring the first section of the paper (formerly heterogeneity) to be more general. I broke out the dimensionality reduction text we had about batch effects into a more general section (Dimensionality reduction and representation learning) that also now contains the representation learning text that was in the prior knowledge section.
Part of the rationale for this move is "reader friendliness" – I expect some readers looking to use more ML in their work will have done some dimensionality reduction.
For reference, the TODOs from that section on #114 are:
<!-- TODO: Use the bits about dimensionality reduction that were in this section originally to talk about dimensionality reduction in general. Readers who are new practicioners might be familiar with or have used PCA - that's ML, too! -->
<!-- TODO: What is dimensionality reduction? -->
In #114, I proposed restructuring the first section of the paper (formerly heterogeneity) to be more general. I broke out the dimensionality reduction text we had about batch effects into a more general section (
Dimensionality reduction and representation learning) that also now contains the representation learning text that was in the prior knowledge section.Part of the rationale for this move is "reader friendliness" – I expect some readers looking to use more ML in their work will have done some dimensionality reduction.
For reference, the TODOs from that section on #114 are:
The changes to the overall introduction to the section formerly known as heterogeneity is tracked here: https://github.com/jaybee84/ml-in-rd/issues/112#issuecomment-692280659