David Brown
-
When is it acceptable to cite your own unpublished internal technical reports?
Paper Discussion
Internal technical reports contain preliminary analysis that has not gone through peer review. Some supervisors encourage citing them while other scholars discourage this practice entirely. What boundary conditions do you apply for citing your own unpublished internal documents?
Justin Scott, Jun Shen and 65 others4 CommentsView more comments-
Only cite internal reports when no equivalent peer‑reviewed alternative exists.
2 -
Make it explicitly clear in‑text that this source remains unpublished internal material.
2 - View more comments
-
-
The hidden cost of chasing SOTA results
Constantly chasing state‑of‑the‑art numbers can consume huge compute and human resources. Sometimes incremental, solid analysis delivers more long‑term scientific value. How do you balance SOTA‑chasing against deeper analytical work?
View more comments-
I try to separate two modes: one for performance improvement, one purely for diagnostic analysis.
2 -
Many SOTA tweaks lack deeper interpretability; you barely understand why gains happen.
2 - View more comments
-
