Efficient Filtered-ANN via Learning-based Query Planning

TL;DR

This paper presents a learning-based query planning framework for filtered approximate nearest neighbor search that dynamically chooses execution strategies, significantly improving speed and recall in vector retrieval tasks.

Contribution

It introduces a novel, lightweight prediction-based framework that optimally selects between pre-filtering and post-filtering for each query, supporting diverse filters and backends.

Findings

Achieves up to 4x acceleration over baselines.

Maintains at least 90% recall in experiments.

Supports various filter types and is dataset-agnostic.

Abstract

Filtered ANN search is an increasingly important problem in vector retrieval, yet systems face a difficult trade-off due to the execution order: Pre-filtering (filtering first, then ANN over the passing subset) requires expensive per-predicate index construction, while post-filtering (ANN first, then filtering candidates) may waste computation and lose recall under low selectivity due to insufficient candidates after filtering. We introduce a learning-based query planning framework that dynamically selects the most effective execution plan for each query, using lightweight predictions derived from dataset and query statistics (e.g., dimensionality, corpus size, distribution features, and predicate statistics). The framework supports diverse filter types, including categorical/keyword and range predicates, and is generic to use any backend ANN index. Experiments show that our method achieves up to 4x acceleration with >= 90% recall comparing to the strong baselines.