GeoKGM: A Multimodal Large Language Model for Zero-Shot Knowledge Graph Completion in Geospatial Databases

Summary: GeoKGM: multimodal LLM framework for zero-shot geospatial KG completion—self-supervised geospatial encoder plus spatial-feature injection into LLMs and multi-task fine-tuning to enable direct inference on unlabeled geospatial DBs. Uses adversarial implicit alignment for cross-domain robustness and outperforms prior SOTA on four real-world datasets. (summarized by gpt-5-mini on Feb 11 2026)

Paper ID: 7582
Venue: SIGMOD
Year: 2026
Pagerank: 5.093636e-05
Overall Rank: 10,374 | 28.83%
DOI: 10.1145/3769796

Incoming Non-self Citations Over Time

No non-self incoming citations found for this paper in this database.

Authors

1. Zhihan Zheng (Beijing Institute of Technology)
2. Haitao Yuan (Nanyang Technological University)
3. Minxiao Chen (Beijing Institute of Technology)
4. Nan Jiang (Nanyang Technological University)
5. Haoning Wang (National University of Singapore)
6. Shangguang Wang (Beijing Institute of Technology)

BibTeX Citation

@inproceedings{zheng_sigmod26,
        title = {{GeoKGM: A Multimodal Large Language Model for Zero-Shot Knowledge Graph Completion in Geospatial Databases}},
        author = {Zheng, Zhihan and Yuan, Haitao and Chen, Minxiao and Jiang, Nan and Wang, Haoning and Wang, Shangguang},
        series = {{SIGMOD} '26},
        booktitle = {Proceedings of the {ACM} {SIGMOD} International Conference on Management of Data},
        publisher = {Association for Computing Machinery},
        doi = {10.1145/3769796},
        url = {https://dl.acm.org/doi/10.1145/3769796},
        year = {2026}
}

Incoming Citations (Sorted by Pagerank)

Showing 0 of 0 citing papers.

Rank	Citing Paper	Year	Venue	Pagerank

Outgoing Citations (Sorted by Pagerank)

Showing 13 of 13 cited papers.

Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.

Rank	Cited Paper	Year	Venue	Pagerank
2,781	Realistic Re-evaluation of Knowledge Graph Completion Methods: An Experimental Study	2020	SIGMOD	8.1281896e-05
3,787	Combining Small Language Models and Large Language Models for Zero-Shot NL2SQL	2024	VLDB	7.1249098e-05
6,341	Cardinality Estimation over Knowledge Graphs with Embeddings and Graph Neural Networks	2024	SIGMOD	5.9068986e-05
7,106	Incremental Tabular Learning on Heterogeneous Feature Space	2023	SIGMOD	5.7005795e-05
8,010	Nuhuo: An Effective Estimation Model for Traffic Speed Histogram Imputation on A Road Network	2024	VLDB	5.5074939e-05
8,077	Deep Transfer Learning for Multi-source Entity Linkage via Domain Adaptation	2022	VLDB	5.4926436e-05
9,061	Mining Geospatial Relationships from Text	2023	SIGMOD	5.3251649e-05
9,062	Shortest Paths Discovery in Uncertain Networks via Transfer Learning	2023	SIGMOD	5.3251649e-05
9,063	Deep Active Alignment of Knowledge Graph Entities and Schemata	2023	SIGMOD	5.3251649e-05
9,071	Data Lakes Empowered by Knowledge Graph Technologies	2021	SIGMOD	5.3251649e-05
10,677	RLER-TTE: An Efficient and Effective Framework for En Route Travel Time Estimation with Reinforcement Learning	2025	SIGMOD	5.093636e-05
10,777	RLOMM: An Efficient and Robust Online Map Matching Framework with Reinforcement Learning	2025	SIGMOD	5.093636e-05
11,452	Automatic Road Extraction with Multi-Source Data Revisited: Completeness, Smoothness and Discrimination	2023	VLDB	5.093636e-05

Semantically Similar Papers

#	Overall Rank	Paper	Year	Venue
1	9,063	Deep Active Alignment of Knowledge Graph Entities and Schemata	2023	SIGMOD
2	8,343	An Interactive Multi-modal Query Answering System with Retrieval-Augmented Large Language Models	2024	VLDB
3	10,023	A Universal Question-Answering Platform for Knowledge Graphs	2023	SIGMOD
4	10,663	Dialogue Benchmark Generation from Knowledge Graphs with Cost-Effective Retrieval-Augmented LLMs	2025	SIGMOD
5	11,045	SDG-KG: A Framework to Compute SDG Indicators with Open Data	2025	VLDB
6	11,226	ZeroEA: A Zero-Training Entity Alignment Framework via Pre-Trained Language Model	2024	VLDB
7	11,061	Large Language Models for Spatial Analysis Queries	2025	VLDB
8	10,819	GeoBloom: Revisiting Lightweight Models for Geographic Information Retrieval	2025	VLDB
9	10,248	Generalized Entity Matching with Adaptivity via Large Language Models	2026	SIGMOD
10	9,061	Mining Geospatial Relationships from Text	2023	SIGMOD