CS Paper Radar

Errpedia: Is Your Exact $k$-NN Vector Search Really Exact?

Yinuo Deng, Ying Zhang, Hanchen Wang, Wenjie Zhang

cs.DB ·submitted 2026-11-17 09:33 UTC · v1
benchmarkdatasetcode/data

Abstract

Exact $k$-nearest-neighbor ($k$-NN) vector search is a common feature of vector libraries and database systems. It serves applications that require exact results and provides the ground truth for evaluating approximate search. However, floating-point rounding and overflow can cause an exact search to return mathematically incorrect top-$k$ result sets. These errors can go unnoticed and affect downstream tasks that rely on exact results. We present Errpedia, a purpose-built dataset for testing the numerical correctness of exact vector search. Errpedia contains 1,616 queries across 21 error categories, with ground truth independently verified using exact rational arithmetic. We evaluate 20 widely used library and database configurations and find that every tested configuration returns mathematically incorrect top-10 result sets on at least 1 applicable query. Failures include misranking nearly equidistant vectors and returning incomplete results: 1 library omits all 10 exact matches to a query, and several systems return fewer than $k$ neighbors. We release Errpedia as a reproducible benchmark for developers and researchers to test their implementations. The dataset is publicly available at https://github.com/dynos01/errpedia.

Closest papers

Nearest neighbours by meaning across the whole index.

Per-Query Recall Guarantees for Approximate Nearest Neighbor Search in Dynamic Vector Databases

cs.DB ·2026-11-05 ·Woong-Kee Loh, Suan Lee
similarity 0.63
Abstract · 66 pages, 15 figures, 19 tables. Extended version of a paper submitted to ICDE 2027

Vector search now runs as an operator inside database systems, where a missed neighbor makes a query answer wrong, yet approximate nearest neighbor (ANN) indexes are tuned only for a target average recall. Tuned to an average recall of 0.99, indexes left 5-19% of in-distribution queries and, without a query log, up to 99.6% of out-of-distribution queries below it (more under updates) and changed a top-5-tag aggregate over the results for up to 27% of queries. In the spirit of approximate query processing, we present PQR-Search, a $k$-ANN operator that guarantees $\Pr[\mathrm{Recall}(q) < τ] \le δ$ for every query and, for $τ< 1$, $\mathbb E[\mathrm{Recall}(q)] \ge τ$. The probability is over a random rotation inside the index alone, so no distributional assumption is needed, only that queries and updates do not depend on it, and the guarantee holds after insertions and deletions without retuning. PQR-Search prunes 1-bit codes of rotated IVF residuals with exact error quantiles and predicts its cost without data-fitted constants. On 20 datasets of up to 100M vectors, failure rates stayed at most $δ$ in every audit, including the NeurIPS'23 streaming runbook, and the aggregate changed in at most 0.04% of trials. The guarantee costs 2.5$\times$ the single-query latency of HNSW tuned as above and 1.3$\times$ lower batch throughput (geometric means). PQR-Search was 26$\times$ faster than FAISS exhaustive search in batch, faster than IVF with $n_{\rm probe}$ chosen afterwards to meet the target on every test query, and 27-76$\times$ faster than PUFFINN, which also guarantees per-query recall.

0.20

Fidelity Synopses: Pre-Execution Plan Selection for Approximate Vector Search

cs.DB cs.IR ·2026-10-30 ·Adithya Ramanatha, Sanskar S. Singh, Maya Ramanath
benchmark similarity 0.62
Abstract

Vector search has become a database operator with many physical implementations: different index families, build-time and search-time parameters, and compressed or disk-resident variants. Each trades cost against fidelity, the fraction of a query's exact top-k nearest neighbors that a configuration returns. Deployments nevertheless choose one configuration offline and use it for every query, so they are sized for their hardest queries and overpay for the rest. On a 5.8M-item LAION benchmark with four embedding models, running each query on the cheapest configuration that suffices for it would roughly halve the cost in a typical setting. Exploiting this requires predicting each configuration's fidelity before running it, which is hard because fidelity is defined by the very exact search that approximate indexes avoid. Relational optimizers solve the analogous problem for cardinality with offline synopses; we propose one for vector search. A fidelity synopsis divides the embedding space into cells around a set of landmarks and records, for every cell and configuration, how often that configuration met a fidelity target on a sample of queries. At query time, one nearest-landmark lookup yields a calibrated estimate for every configuration, and an optimizer chooses configurations so that the workload misses the target no more often than its requirement allows. We test the two premises the synopsis relies on, that fidelity varies smoothly across the embedding space and that its loss is specific to each configuration. Held to the same requirement a static deployment is built to meet, synopsis-guided selection lowers cost in all 16 combinations of model and query direction, to a median of 0.72x the static deployment, with a batched lookup of 21-32 microseconds per query. The saving persists against the best static choice among the same indexes, and within a single index at no extra memory.

0.09

Index-Free Dynamic Edge Retrieval with Energy-Tail-Aware Partial Scans

cs.IR ·2026-09-01 ·Mohammad Arif Rasyidi, Omar Alhussein
publishedcode/data similarity 0.61
Abstract · Accepted for presentation at the 2026 IEEE 12th World Forum on Internet of Things (WF-IoT)

Dynamic maximum inner-product search (MIPS) returns the $K$ stored vectors with the largest dot products with a query while allowing the dataset to change through insertions, replacements, and deletions. For edge retrieval, the challenge is to achieve high recall and fast queries without making updates expensive. Full-vector scanning keeps updates simple but compares each query with every stored vector, while indexed methods reduce query cost at the expense of maintaining additional structures during updates. We propose ETAR, an index-free method that reduces query work while preserving simple updates. ETAR keeps the query coordinates with the largest squared values until they cover most of its total squared magnitude and treats the rest as a low-magnitude tail. It estimates similarity from the retained coordinates using a compact lower-precision representation, corrects for skipped coordinates, and reranks a fixed number of candidates using full-precision vectors. Across five runs on nine static datasets, ETAR averages 99.2% Recall@10, the fraction of exact top-10 results recovered, while running over 4$\times$ faster than exact scanning at a representative setting. This speedup also extends to an ARM-based mobile device, where ETAR is up to 6.9$\times$ faster across four synthetic distributions. Under five streaming workloads, it maintains 100% Recall@10 at every measured point without index rebuilds. Overall, ETAR offers a practical middle ground for dynamic MIPS by reducing query cost while retaining simple, index-free updates. Code is available at https://github.com/arasyi/etar-mips.

0.47

CapCover Filter: Semantic Approximate Membership Queries over Embeddings

cs.DB cs.AI ·2026-11-15 ·Dvir David Biton, Shay Hahiashvili, Nadav Keren, Gil Einziger, Roy Friedman
similarity 0.61
Abstract

Determining whether an embedding is sufficiently similar to any vector in a vector store without performing a nearest-neighbor search can substantially improve query throughput in vector-store systems. We study Semantic Approximate Membership Queries (SAMQs), which allow queries to be rejected when no indexed vector meets the requested similarity threshold, while guaranteeing no false negatives. We introduce CapCover Filter, a geometric data structure that summarizes a set of embeddings using spherical caps. Its representation is independent of the query threshold, allowing a single filter to support arbitrary thresholds without reconstruction. We evaluate alternative construction and quantization methods, compare CapCover Filter with competing filters, and measure its impact on query throughput. Across five real-world datasets, at a cosine similarity threshold of 0.95, CapCover Filter eliminates approximately 20--100\% of index accesses while using only 2\% of the raw (FP32) embedding memory. When placed before a FAISS flat index, it improves query throughput by up to 382\% and by approximately 116\% on average across the five datasets, while returning the same range-search result. We also evaluate lower similarity thresholds. Our results show that CapCover Filter can substantially improve vector-store throughput when similarity thresholds are tight, retrieval is expensive, and indexed embeddings are highly clustered. We release an optimized implementation of CapCover Filter and its evaluation framework.

0.14

A Comprehensive Empirical Evaluation of Vector Database Systems for Approximate Nearest Neighbor Search: Performance, Quality, and Resource Trade-offs

cs.IR ·2026-08-13 ·Ashen Rashmiks, Tiroshan Madushanka
code/data similarity 0.60
Abstract

Vector databases have emerged as critical infrastructure for modern artificial intelligence applications, particularly retrieval-augmented generation (RAG), semantic search, and recommendation systems. Despite their growing importance, there remains a significant gap in comprehensive, reproducible benchmarks that jointly evaluate retrieval quality, query latency, throughput, and resource utilization. We present a systematic empirical evaluation of seven prominent vector database systems: FAISS, Qdrant, Milvus, Weaviate, Chroma, pgvector, and LanceDB. Our methodology spans six diverse datasets, from classical computer-vision descriptors (SIFT, GIST) to transformer-based text embeddings (MS MARCO, GloVe), encompassing over 4 million vectors at dimensionalities from 96 to 960. We measure 15 metrics spanning retrieval quality (Recall@K, Precision@K, MRR, NDCG@K, Hit Rate@K), query performance (latency percentiles, QPS, cold-start latency), and resource consumption (index build time, memory, storage). On SIFT1M, FAISS achieves the highest single-node throughput (866 QPS) but lacks database operational features; Weaviate provides the best out-of-the-box recall (> 99%); Qdrant offers the best latency among full databases (4.55~ms median); and LanceDB trades retrieval quality for substantially faster index construction. We derive system-selection guidelines for practitioners and release our benchmarking framework as open-source software.

0.20

GINGER: Spicing Up Graph-Based Vector Search

cs.IR cs.DC ·2026-11-05 ·Daniel Salwasser, Lars Gottesbüren, Laxman Dhulipala, Jakub Łącki, Guy Blelloch +5 more
code/data similarity 0.60
Abstract

We present \textsc{Ginger}, a state-of-the-art open-source C++ library for graph-based approximate nearest neighbor search (ANNS). We developed \textsc{Ginger} at Google to push the speed and memory efficiency of graph search on large and high-dimensional embedding datasets. \textsc{Ginger} surpasses existing ANNS libraries with regards to query throughput and indexing time while matching the memory footprint of the most space-efficient competitors. \textsc{Ginger} builds upon recent advances in proximity graph construction (PiPNN) and vector quantization (TurboQuant). We introduce Nested TurboQuant (NTQ), a two-level quantizer which supports low precision graph navigation and high precision reranking in a single code. This representation achieves near floating point precision with $2.6\times$ less memory. We also compress the size of our proximity graphs by up to $2.5\times$ without impacting query throughput. We combine graph reordering with a new lossless compression codec called AdaptiveGap, which builds upon StreamVByte. At $95\%$ recall, \textsc{Ginger} speeds up query throughput by an average of $1.8\times$ over each dataset's fastest competitor and by at least $2.3\times$ over any single library. This lead grows to $2.9\times$ over compact libraries on high-dimensional embeddings.

0.26

KatoANN: Enabling Near-Exact Vector Search at the Speed of Lossy Scans with CPU Matrix Engines

cs.DB cs.AR ·2026-10-20 ·Soobin Cho, Gyuyeong Kim
similarity 0.59
Abstract

Distance scans in partition-family approximate nearest-neighbor (ANN) search are typically limited by memory bandwidth. Existing systems therefore improve throughput by reading fewer bytes per candidate, often at the cost of recall. We present KatoANN, a CPU ANN search system that removes this tradeoff by batching distance scans on Intel AMX, a CPU matrix engine. Batching allows multiple queries to reuse each fetched vector code, shifting the scan from memory-bound to compute-bound and enabling near-exact distances at the speed of a lossy scan. KatoANN stores each vector once as a 16-bit SQ16 code split into high- and low-byte planes. Its adaptive rank-then-refine scan first ranks all candidates using only the high byte, then refines a small survivor set using the full-byte code. Both stages use the same representation, avoiding a second copy of the raw vectors. Experiments across diverse high-dimensional workloads show that KatoANN is the fastest partition-family ANN search system in our evaluation throughout the high-recall range.

0.24

CertEx: Exact Visual Retrieval at Scale by Certified Exclusion

cs.CV ·2026-10-15 ·Dongfang Zhao
theory similarity 0.59
Abstract · 30 pages, 12 figures

Exact $k$-nearest-neighbor (kNN) search over image descriptors and embeddings underlies visual retrieval, near-duplicate detection, the evaluation of learned representations, and increasingly the memory of LLM agents. However, exhaustive search computes every distance at full precision, whereas approximate indexes return recall below one. We observe that exact search only needs a proof that each excluded vector cannot be a neighbor. Because such a proof only has to be conservative, it can be checked in low precision, much as the retina examines the whole scene at low resolution. We call this principle certified exclusion and realize it in an exact kNN method for GPUs, which we call CertEx. Specifically, CertEx stores a lower-bounding projection of each vector in half precision and folds a rounding certificate into the same tensor-core dot product, so that only the few vectors that the test keeps need exact distances. We prove that CertEx returns the exact kNN of every query. On sixteen datasets with up to 100M vectors, CertEx is exact for every query at 7.1$\times$ to 18.0$\times$ the throughput of the fastest cuVS search on one H100 (8.8$\times$ to 11.8$\times$ on an A100) and outperforms quantized and graph indexes whenever every query must be exact.

0.43