Lower Bounds on Performance of Metric Tree Indexing Schemes for Exact Similarity Search in High Dimensions

TL;DR

This paper establishes fundamental lower bounds on the efficiency of metric tree indexing schemes for exact similarity search in high-dimensional spaces, demonstrating the curse of dimensionality under certain mathematical assumptions.

Contribution

It provides a rigorous theoretical analysis showing that hierarchical metric-tree schemes cannot outperform superpolynomial time complexity in high dimensions, under specified conditions.

Findings

Lower bound of (n^{1/4}) on expected search performance

Superpolynomial complexity in the intrinsic dimension d

Analysis based on VC dimension and Lipschitz functions

Abstract

Within a mathematically rigorous model, we analyse the curse of dimensionality for deterministic exact similarity search in the context of popular indexing schemes: metric trees. The datasets $X$ are sampled randomly from a domain $\Omega$, equipped with a distance, $\rho$, and an underlying probability distribution, $\mu$. While performing an asymptotic analysis, we send the intrinsic dimension $d$ of $\Omega$ to infinity, and assume that the size of a dataset, $n$, grows superpolynomially yet subexponentially in $d$. Exact similarity search refers to finding the nearest neighbour in the dataset $X$ to a query point $\omega\in\Omega$, where the query points are subject to the same probability distribution $\mu$ as datapoints. Let $\mathscr F$ denote a class of all 1-Lipschitz functions on $\Omega$ that can be used as decision functions in constructing a hierarchical metric tree indexing scheme. Suppose the VC dimension of the class of all sets $\{\omega\colon f(\omega)\geq a\}$, $a\in\R$ is $o(n^{1/4}/\log^2n)$. (In view of a 1995 result of Goldberg and Jerrum, even a stronger complexity assumption $d^{O(1)}$ is reasonable.) We deduce the $\Omega(n^{1/4})$ lower bound on the expected average case performance of hierarchical metric-tree based indexing schemes for exact similarity search in $(\Omega,X)$. In paricular, this bound is superpolynomial in $d$.