Hacker News new | ask | show | jobs
by minimaxir 893 days ago
Both the viz, and that the 2D UMAP projection is actually enough to get accurately delineated topics.

Hence why I think the typical embedding dimensionality is way way too high.
2 comments
markisus 893 days ago
Do you think 1D could work? Maybe topic-space is some sort of tree-shaped structure where documents live in the thin strands.
link
minimaxir 893 days ago
1D could work on certain datasets but it wouldn't be ideal.
link
adammarples 893 days ago
Why not just embed directly to 2d? Does it give worse results than UMAP?
link