SAGA: Synthetic Agentic Graph Architecture for Temporal Benchmark Generation

2026-07-19

AI Summary

Existing benchmarks for temporal graphs suffer from data scarcity due to privacy constraints and high annotation costs. This work proposes the first four-stage generation framework that prioritizes structure before semantics: it begins by constructing a power-law backbone, then employs causally aware temporal chunking with parallel scheduling, injects multi-domain semantics via RAG-enhanced LLM agents, and finally uses a temporal state alignment engine to automatically replay and generate realistic anomaly labels. The approach holistically ensures structural fidelity, semantic richness, and anomaly interpretability, overcoming limitations of conventional generators and pure LLM-based methods. On a single H100 GPU, the framework generates 500,000 temporally labeled edges with controllable anomalies in under 90 minutes—supporting graphs with up to 100,000 nodes and clustering coefficients exceeding 0.99—and enables real-time visualization and interactive refinement.

Abstract

High quality temporal graph benchmarks with rich semantics and ground-truth anomaly labels are essential for training graph neural networks, yet remain scarce due to privacy constraints and annotation costs. We present SAGA (Synthetic Agentic Graph Architecture), a system for generating large-scale, semantically rich temporal graphs via a four-phase pipeline. Our Skeleton-First, Semantics-Second architecture decouples structure from semantics: (S) an O(1)-per-edge skeleton generator produces power-law graphs; (A) a dispatcher partitions causally ordered time blocks for parallel execution; (G) LLM agents inject domain semantics using RAG-based rule bases across four domains; and (A) a state alignment engine resolves conflicts via temporal replay, yielding anomaly labels as natural byproducts. Unlike structural generators (e.g., LDBC SNB, Kronecker/R-MAT) or purely LLM-based approaches, SAGA achieves structural realism, semantic richness, and automatic anomaly labeling in a unified framework. On a single H100 GPU with vLLM batching, SAGA generates 500,000 temporal edges with controlled anomalies in under 90 minutes, scaling to 100,000 nodes while maintaining clustering coefficients above 0.99. The system supports real-time pipeline visualization, interactive multi-domain tuning (Finance/AML, Network/IDS, Cyber/APT, Transportation), and a CLI for large-scale GPU-based experiments.

Problem

• temporal graph benchmarks
• semantic richness
• anomaly labeling
• data scarcity
• privacy constraints

Innovation

• Temporal Graph Generation
• Synthetic Benchmark
• LLM Agents
• Anomaly Labeling
• Skeleton-First Semantics-Second