{"paper":{"title":"Neuromorphic Graph Anomaly Detection via Adaptive STDP and Spiking Graph Neural Networks","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"A spiking neural network framework called ASTDP-GAD detects anomalies in dynamic graphs using adaptive STDP learning for neuromorphic energy efficiency.","cross_cats":["cs.LG"],"primary_cat":"cs.NE","authors_text":"Abdul Joseph Fofanah, David Chen, Kwabena Sarpong, Lian Wen, Tsungcheng Yao","submitted_at":"2026-04-29T04:45:11Z","abstract_excerpt":"Anomaly detection in dynamic networks is critical for applications from cybersecurity to industrial monitoring, yet existing methods face challenges in energy efficiency, temporal precision, and adaptability. This paper introduces ASTDP-GAD, a novel Adaptive Spiking Temporal Dynamics Plasticity framework for Graph Anomaly Detection that integrates spiking graph neural networks with STDP learning for energy-efficient neuromorphic detection in dynamic networks. Our framework unifies spiking neural computation, STDP learning, and graph-based anomaly detection through the following key innovations"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Theoretical analysis provides rigorous guarantees: spike encoding preserves input information with resolution scaling linearly in simulation steps and hidden dimension; LIFGAT approximates any continuous attention function; hypergraph memory converges to optimal prototypes; contrast pooling achieves provable anomaly selection bounds; STDP learning converges stably; and multi-factor fusion produces calibrated scores with up to 5× variance reduction.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The assumption that the listed theoretical approximations and convergence results transfer to real-world dynamic graph data without significant degradation from discretization, noise, or hyperparameter choices in the adaptive LIF and STDP components.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"ASTDP-GAD unifies spiking neural computation, STDP learning, and graph anomaly detection with claimed theoretical guarantees on encoding, convergence, and score calibration.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"A spiking neural network framework called ASTDP-GAD detects anomalies in dynamic graphs using adaptive STDP learning for neuromorphic energy efficiency.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"49c80ef84266636aa8cdf29c485c424472d2eaed42f6c11abaec229b5317b370"},"source":{"id":"2605.13863","kind":"arxiv","version":1},"verdict":{"id":"d6f30d75-9137-4e05-b8ad-c241a9969ae5","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-15T07:15:03.212897Z","strongest_claim":"Theoretical analysis provides rigorous guarantees: spike encoding preserves input information with resolution scaling linearly in simulation steps and hidden dimension; LIFGAT approximates any continuous attention function; hypergraph memory converges to optimal prototypes; contrast pooling achieves provable anomaly selection bounds; STDP learning converges stably; and multi-factor fusion produces calibrated scores with up to 5× variance reduction.","one_line_summary":"ASTDP-GAD unifies spiking neural computation, STDP learning, and graph anomaly detection with claimed theoretical guarantees on encoding, convergence, and score calibration.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The assumption that the listed theoretical approximations and convergence results transfer to real-world dynamic graph data without significant degradation from discretization, noise, or hyperparameter choices in the adaptive LIF and STDP components.","pith_extraction_headline":"A spiking neural network framework called ASTDP-GAD detects anomalies in dynamic graphs using adaptive STDP learning for neuromorphic energy efficiency."},"references":{"count":79,"sample":[{"doi":"","year":2025,"title":"A generalizable anomaly detection method in dynamic graphs","work_id":"a2a28211-3f73-4086-a607-4a8a5c4584f7","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"Graph anomaly detection in time series: A survey","work_id":"1c3b1683-4942-41a3-ab69-fc602cb6b996","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"Deep anomaly detection on attributed networks by graph update: Y","work_id":"a1bab462-4bae-40a3-877f-53ff8a7b8e28","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"Ge-gnn: Gated edge-augmented graph neural network for fraud detection","work_id":"1b148d6d-2076-48ea-bc1b-7664679ecce8","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"Gat-ad: Graph attention networks for contextual anomaly detection in network monitoring","work_id":"e9a77515-37a1-4a21-a7cf-beb1b31b4909","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":79,"snapshot_sha256":"411ef9b9b059bae95fc3e46a129a466d7868c7e30f550322ead5760224906309","internal_anchors":1},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}