About Margin Notes
A notebook.
Many perspectives.
Reading notes on statistics and machine learning, organized around the questions that connect papers.
Follow a question
The collection brings papers by different authors into a shared reading space. Current themes include repeated computation and parameter sharing, predictive uncertainty, and the practice of comparing and reproducing research. Inclusion is a starting point for study, not a ranking.
Keep the evidence attached
Each paper note retains the original title, authors, source, and the version being discussed. Reported findings are separated from explanations and suggested reading questions. Reading guides offer general methods for inspecting research. For a scientific claim, consult and cite the original work.
A site centered on the work
Margin Notes does not publish its operator’s personal profile or claim an institutional affiliation. Paper bylines identify the research being discussed; they are not bylines for the site’s operator. Notes are updated when there is a substantive correction, a new source, or a useful connection to add.
Return to the notes