Towards Neural Graph Data Management

2026-02-27

AI Summary

Real-world data is inherently heterogeneous, dynamically evolving, and often incomplete, posing significant challenges for existing data management systems to effectively uncover implicit structures and perform robust reasoning under noise and continuous updates. To address this gap, this work proposes NGDBench, the first unified benchmark for neural graph data management. NGDBench pairs clean latent graphs with perturbed observed graphs, integrating structured and unstructured data while supporting Cypher queries, dynamic graph operations, and emerging neural query techniques such as Text-to-Cypher and GraphRAG. Experimental results demonstrate that current neural query methods exhibit notable fragility in noisy and dynamic settings, underscoring the critical role of NGDBench in advancing the development of next-generation neural databases that are both robust and state-traceable.

Abstract

While AI systems have made remarkable progress in processing unstructured text, structured data such as graphs stored in databases, continues to grow rapidly yet remains difficult for neural models to effectively utilize. We introduce NGDBench, a unified benchmark for evaluating neural graph database capabilities across five diverse domains, including finance, medicine, and AI agent tooling. Unlike prior benchmarks limited to elementary logical operations, NGDBench supports the full Cypher query language, enabling complex pattern matching, variable-length paths, and numerical aggregations, while incorporating realistic noise injection and dynamic data management operations. Our evaluation of state-of-the-art LLMs and RAG methods reveals significant limitations in structured reasoning, noise robustness, and analytical precision, establishing NGDBench as a critical testbed for advancing neural graph data management. Our code and data are available at https://github.com/HKUST-KnowComp/NGDBench.

Problem

• neural data management
• graph data
• noise robustness
• dynamic state tracking
• latent relationships

Innovation

• Neural Data Management
• Graph Benchmark
• Latent Graph Reasoning
• Dynamic Data Updates
• Noise-Robust Querying