jl-lemma-bound-tight-up-to-constants
IN premise — summaries/2026/08/24/wiki-JohnsonE28093Lindenstrauss_lemma-chunk-1.md
Created 2026-08-24T17:11:13+00:00
The Johnson-Lindenstrauss dimension requirement is tight up to a constant factor: there exist N-point sets in ℝⁿ requiring Ω(log N / ε²) dimensions for any (1±ε) bi-Lipschitz embedding.
Summary
The Johnson-Lindenstrauss bound is not just a convenient trick; it is essentially the best you can do. There are configurations of points that genuinely cannot be embedded into a much smaller space while preserving pairwise distances, so any system relying on dimensionality reduction should treat the log-N / epsilon-squared threshold as a hard wall rather than something that might be beaten by a smarter algorithm.