QASP: Query-Adaptive Robust Vector Search Policy

2026-07-31

AI Summary

This work addresses the limitations of existing vector retrieval systems that employ fixed search parameters, which lead to substantial inter-query recall variability and mask performance fluctuations under conventional average recall metrics. The paper proposes the first generalizable query-adaptive search strategy: it predicts the full recall progress curve for each query via a lightweight supervised regression model, then dynamically adjusts search depth by leveraging scale-invariant features and a reactive refinement procedure to align with a target recall level. Requiring only a single training phase, the method generalizes across diverse recall targets and index configurations without retraining, offers finite-sample guarantees, and achieves near-theoretical-optimal loss. Evaluated on large-scale datasets and hierarchical indexes, it attains 99% recall while reducing data accesses by 80%, significantly lowering both recall variance and target deviation.

Abstract

A fundamental challenge of vector search is achieving consistently high recall while minimizing computational costs. Fixed search parameters cause significant performance variance across queries, and conventional evaluation on average recall masks these per-query disparities. We introduce QASP (Query-Adaptive robust vector Search Policy), which predicts the complete recall progression curve per query via a single upfront supervised regression, from which a search policy is derived for any recall target; this avoids iterative model invocations during search or separate predictors per target. By predicting normalized recall values with scale-invariant features and pre-search inference, QASP generalizes across recall targets, index configurations, and datasets. Its fine-grained progress predictions further enable a lightweight reactive complement that adjusts search depth based on predicted-versus-observed deviations without additional inference. We prove that QASP requires a finite training sample independent of dataset size and dimensionality, that its loss exceeds the irreducible lower bound of any fixed policy by a vanishing margin, and that its data access savings over fixed probing grow exponentially in intrinsic dimensionality. Experimentally, QASP achieves significantly lower recall variance and deviation from target, higher query satisfaction rate, and scales to large data and hierarchical indices without retraining, achieving 99% recall with 80% less data access.

Problem

• vector search
• recall variance
• query-adaptive
• computational cost
• search policy

Innovation

• query-adaptive search
• recall progression prediction
• scale-invariant features
• pre-search inference
• vector search optimization