13
0

Optimality of the Johnson-Lindenstrauss Dimensionality Reduction for Practical Measures

Abstract

It is well known that the Johnson-Lindenstrauss dimensionality reduction method is optimal for worst case distortion. While in practice many other methods and heuristics are used, not much is known in terms of bounds on their performance. The question of whether the JL method is optimal for practical measures of distortion was recently raised in BFN19 (NeurIPS'19). They provided upper bounds on its quality for a wide range of practical measures and showed that indeed these are best possible in many cases. Yet, some of the most important cases, including the fundamental case of average distortion were left open. In particular, they show that the JL transform has 1+ϵ1+\epsilon average distortion for embedding into kk-dimensional Euclidean space, where k=O(1/ϵ2)k=O(1/\epsilon^2), and for more general qq-norms of distortion, k=O(max{1/ϵ2,q/ϵ})k = O(\max\{1/\epsilon^2,q/\epsilon\}), whereas tight lower bounds were established only for large values of qq via reduction to the worst case. In this paper we prove that these bounds are best possible for any dimensionality reduction method, for any 1qO(log(2ϵ2n)ϵ)1 \leq q \leq O(\frac{\log (2\epsilon^2 n)}{\epsilon}) and ϵ1n\epsilon \geq \frac{1}{\sqrt{n}}, where nn is the size of the subset of Euclidean space. Our results imply that the JL method is optimal for various distortion measures commonly used in practice such as stress, energy and relative error. We prove that if any of these measures is bounded by ϵ\epsilon then k=Ω(1/ϵ2)k=\Omega(1/\epsilon^2) for any ϵ1n\epsilon \geq \frac{1}{\sqrt{n}}, matching the upper bounds of BFN19 and extending their tightness results for the full range moment analysis. Our results may indicate that the JL dimensionality reduction method should be considered more often in practical applications, and the bounds we provide for its quality should be served as a measure for comparison when evaluating the performance of other methods and heuristics.

View on arXiv
Comments on this paper