pith. sign in

Therefore, the KL divergence approaches0whenσ=ω(k −1/4)

Using the above formula, we get 2DKL(R1||R2) =klog σ2 + (1/k) σ2 −k+k σ2 σ2 + (1/k) +klog σ2 σ2 + (1/k) −k+k 1 + (1/kσ2) =−2k+k 1 1 + (1/kσ2) +k+ 1 σ2 ≤ 1 kσ4 , using1/(1+η)≤1−η+η 2 forη∈[0,1] · 2023

1 Pith paper cite this work. Polarity classification is still indexing.

1 Pith paper citing it
browse 1 citing papers

fields

cs.DS 1

years

2026 1

verdicts

UNVERDICTED 1

representative citing papers

SVD Provably Denoises Nearest Neighbor Data

cs.DS · 2026-04-04 · unverdicted · novelty 7.0

SVD provably recovers the uncorrupted nearest neighbor from noisy data when σ is O(1/k^{1/4}), with a matching lower bound showing the threshold is necessary.

citing papers explorer

Showing 1 of 1 citing paper.

  • SVD Provably Denoises Nearest Neighbor Data cs.DS · 2026-04-04 · unverdicted · none · ref 8

    SVD provably recovers the uncorrupted nearest neighbor from noisy data when σ is O(1/k^{1/4}), with a matching lower bound showing the threshold is necessary.