SeedER: Seed-and-Expand Retrieval from Knowledge Graphs

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Shirzad, Hamed, Wenkel, Frederik, Beaini, Dominique, Sutherland, Danica J., Noutahi, Emmanuel
Format:	Preprint
Published:	2026
Subjects:	Machine Learning
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

_version_	1866913156864933888
author	Shirzad, Hamed Wenkel, Frederik Beaini, Dominique Sutherland, Danica J. Noutahi, Emmanuel
author_facet	Shirzad, Hamed Wenkel, Frederik Beaini, Dominique Sutherland, Danica J. Noutahi, Emmanuel
contents	Knowledge graphs (KGs) offer a rich representation for relational knowledge, but their irregular structure makes retrieval challenging: ego-graph expansion grows rapidly, and dense embedding methods struggle with multi-hop compositional queries. Existing agent-based graph exploration approaches, while expressive, are often too expensive for large-scale retrieval. We introduce SeedER (Seed-and-Expand Retrieval), a retrieval framework that explicitly leverages KG structure through iterative, low-cost expansion. SeedER first seeds a compact set of core nodes using lightweight dense and entity-based retrieval, then selectively expands this set via a learned graph-aware policy trained with reinforcement learning. This design decomposes global reasoning into reusable local decisions, enabling efficient discovery of query-relevant nodes while tightly controlling expansion cost. We show theoretical limitations of dense retrieval on compositional graph queries, and establish advantages of SeedER from both compositional generalization and graph-constrained submodular optimization perspectives. Empirically, SeedER substantially improves recall with compact candidate sets over strong dense and graph-augmented baselines, making it an effective first-stage retriever for knowledge-intensive reasoning systems.
format	Preprint
id	arxiv_https___arxiv_org_abs_2605_23753
institution	arXiv
publishDate	2026
record_format	arxiv
spellingShingle	SeedER: Seed-and-Expand Retrieval from Knowledge Graphs Shirzad, Hamed Wenkel, Frederik Beaini, Dominique Sutherland, Danica J. Noutahi, Emmanuel Machine Learning Knowledge graphs (KGs) offer a rich representation for relational knowledge, but their irregular structure makes retrieval challenging: ego-graph expansion grows rapidly, and dense embedding methods struggle with multi-hop compositional queries. Existing agent-based graph exploration approaches, while expressive, are often too expensive for large-scale retrieval. We introduce SeedER (Seed-and-Expand Retrieval), a retrieval framework that explicitly leverages KG structure through iterative, low-cost expansion. SeedER first seeds a compact set of core nodes using lightweight dense and entity-based retrieval, then selectively expands this set via a learned graph-aware policy trained with reinforcement learning. This design decomposes global reasoning into reusable local decisions, enabling efficient discovery of query-relevant nodes while tightly controlling expansion cost. We show theoretical limitations of dense retrieval on compositional graph queries, and establish advantages of SeedER from both compositional generalization and graph-constrained submodular optimization perspectives. Empirically, SeedER substantially improves recall with compact candidate sets over strong dense and graph-augmented baselines, making it an effective first-stage retriever for knowledge-intensive reasoning systems.
title	SeedER: Seed-and-Expand Retrieval from Knowledge Graphs
topic	Machine Learning
url	https://arxiv.org/abs/2605.23753