Fast and Efficient Approximate Nearest Neighbor Search for High-Dimensional LLM Embeddings

Nico Hezel; Kai Uwe Barthel; Bruno Schilling; Konstantin Schall; Andre Moelle; Klaus Jung

Fast and Efficient Approximate Nearest Neighbor Search for High-Dimensional LLM Embeddings

Information Retrieval 2026-07-23 v1

Authors: Nico Hezel , Kai Uwe Barthel , Bruno Schilling , Konstantin Schall , Andre Moelle , Klaus Jung

Abstract

The annual SISAP Indexing Challenge benchmarks Approximate Nearest Neighbor Search (ANNS) algorithms under rigorous constraints. This paper presents our submissions for the 2026 edition, addressing both $k$ -Nearest Neighbor Graph (kNNG) construction on 1024-dimensional BGE-M3 embeddings (Task 1) and Maximum Inner Product Search (MIPS) on unnormalized Llama-3.2-8B features (Task 2). To optimize construction speed, we utilize Equi-Voronoi Polytopes (EVP) for efficient quantization, supplemented by targeted reranking strategies to maintain high recall. For MIPS, we transform the asymmetric inner product problem into a Euclidean search space via dimensionality augmentation. To reduce query latency and optimize memory access, we introduce a 1D presorting mechanism via Fast Linear Assignment Sorting (FLAS) prior to graph construction. This significantly improves spatial locality and cache hit rates during subsequent graph traversal. Source Code: https://github.com/Visual-Computing/sisap26-deglib

Cite

@article{arxiv.2607.20957,
  title  = {Fast and Efficient Approximate Nearest Neighbor Search for High-Dimensional LLM Embeddings},
  author = {Nico Hezel and Kai Uwe Barthel and Bruno Schilling and Konstantin Schall and Andre Moelle and Klaus Jung},
  journal= {arXiv preprint arXiv:2607.20957},
  year   = {2026}
}