FuseSampleAgg: One-Pass Neighborhood Estimation for Budgeted Knowledge-Graph Refresh and Validation
arXiv:2511.13645v2 Announce Type: replace Abstract: Operational knowledge-graph (KG) pipelines in networking and cybersecurity increasingly need to refresh embeddings under strict time, memory, and audit budgets, especially as curated feeds and LLM-assisted extraction accelerate KG updates. A recurring per-step cost in mini-batch KG learning is neighborhood-context estimation: uniform neighbor sampling without replacement followed by mean aggregation. Common frameworks implement this estimator through sampled-subgraph materialization and intermediate feature gathers, adding kernel launches, al
The increasing complexity and rapid refresh demands of operational knowledge graphs, particularly with LLM integration, necessitate more efficient and scalable computational methods.
This development addresses a critical bottleneck in the performance and cost-effectiveness of AI systems relying on dynamic knowledge graphs, impacting their scalability and real-world applicability.
A more efficient method for neighborhood estimation in knowledge graphs allows for faster, cheaper, and more auditable updates to AI models, directly influencing the agility of AI-driven applications.
- · AI/ML researchers and developers
- · Cloud service providers
- · Cybersecurity and networking industries
- · Companies using LLM-assisted data extraction
- · Inefficient knowledge graph management systems
- · High-latency operational AI systems
Reduced computational cost and time for updating knowledge graph embeddings in large-scale AI applications.
Accelerated development and deployment of real-time AI agents and intelligent systems that rely on fresh, accurate data.
Enhanced operational robustness and auditability of AI systems across critical infrastructure, potentially increasing trust and adoption.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.LG