{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:ROOI4P2EUYHAAWVMNPVPUUDKQC","short_pith_number":"pith:ROOI4P2E","schema_version":"1.0","canonical_sha256":"8b9c8e3f44a60e005aac6beafa506a809cf573d62b6e00c9be61b4b98c075107","source":{"kind":"arxiv","id":"2506.17709","version":1},"attestation_state":"computed","paper":{"title":"CEGA: A Cost-Effective Approach for Graph-Based Model Extraction and Acquisition","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CR","stat.ML"],"primary_cat":"cs.LG","authors_text":"Bolin Shen, Ken Anderson, Menghan Lin, Molei Liu, Tianxi Cai, Yushun Dong, Zebin Wang","submitted_at":"2025-06-21T13:11:42Z","abstract_excerpt":"Graph Neural Networks (GNNs) have demonstrated remarkable utility across diverse applications, and their growing complexity has made Machine Learning as a Service (MLaaS) a viable platform for scalable deployment. However, this accessibility also exposes GNN to serious security threats, most notably model extraction attacks (MEAs), in which adversaries strategically query a deployed model to construct a high-fidelity replica. In this work, we evaluate the vulnerability of GNNs to MEAs and explore their potential for cost-effective model acquisition in non-adversarial research settings. Importa"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2506.17709","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-06-21T13:11:42Z","cross_cats_sorted":["cs.CR","stat.ML"],"title_canon_sha256":"d456a5dfa7dc23f947819221f9904cf92ad0516985347d35037d036814ce6b3c","abstract_canon_sha256":"63630f505f660d7ff0ead17fdf05e2f4ec26b585cf539ef8540fde5363986784"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:33:25.031386Z","signature_b64":"3IY7D72CBITZU5qntSAGQZcRMe13DqBQMlpQeS8SqJSqlApJIWBbC8/KT1bQ+/W6WsxPWJKy8+q0lIGPl8vsCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8b9c8e3f44a60e005aac6beafa506a809cf573d62b6e00c9be61b4b98c075107","last_reissued_at":"2026-07-05T11:33:25.030863Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:33:25.030863Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"CEGA: A Cost-Effective Approach for Graph-Based Model Extraction and Acquisition","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CR","stat.ML"],"primary_cat":"cs.LG","authors_text":"Bolin Shen, Ken Anderson, Menghan Lin, Molei Liu, Tianxi Cai, Yushun Dong, Zebin Wang","submitted_at":"2025-06-21T13:11:42Z","abstract_excerpt":"Graph Neural Networks (GNNs) have demonstrated remarkable utility across diverse applications, and their growing complexity has made Machine Learning as a Service (MLaaS) a viable platform for scalable deployment. However, this accessibility also exposes GNN to serious security threats, most notably model extraction attacks (MEAs), in which adversaries strategically query a deployed model to construct a high-fidelity replica. In this work, we evaluate the vulnerability of GNNs to MEAs and explore their potential for cost-effective model acquisition in non-adversarial research settings. Importa"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.17709","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2506.17709/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2506.17709","created_at":"2026-07-05T11:33:25.030924+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.17709v1","created_at":"2026-07-05T11:33:25.030924+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.17709","created_at":"2026-07-05T11:33:25.030924+00:00"},{"alias_kind":"pith_short_12","alias_value":"ROOI4P2EUYHA","created_at":"2026-07-05T11:33:25.030924+00:00"},{"alias_kind":"pith_short_16","alias_value":"ROOI4P2EUYHAAWVM","created_at":"2026-07-05T11:33:25.030924+00:00"},{"alias_kind":"pith_short_8","alias_value":"ROOI4P2E","created_at":"2026-07-05T11:33:25.030924+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":4,"sample":[{"citing_arxiv_id":"2606.19185","citing_title":"AGDN: Learning to Solve Traveling Salesman Problem with Anisotropic Graph Diffusion Network","ref_index":60,"is_internal_anchor":true},{"citing_arxiv_id":"2605.12827","citing_title":"GraphIP-Bench: How Hard Is It to Steal a Graph Neural Network, and Can We Stop It?","ref_index":27,"is_internal_anchor":true},{"citing_arxiv_id":"2605.30470","citing_title":"Can Subgraph Explanations Be Weaponized to Steal Graph Neural Networks?","ref_index":40,"is_internal_anchor":true},{"citing_arxiv_id":"2605.12827","citing_title":"GraphIP-Bench: How Hard Is It to Steal a Graph Neural Network, and Can We Stop It?","ref_index":27,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ROOI4P2EUYHAAWVMNPVPUUDKQC","json":"https://pith.science/pith/ROOI4P2EUYHAAWVMNPVPUUDKQC.json","graph_json":"https://pith.science/api/pith-number/ROOI4P2EUYHAAWVMNPVPUUDKQC/graph.json","events_json":"https://pith.science/api/pith-number/ROOI4P2EUYHAAWVMNPVPUUDKQC/events.json","paper":"https://pith.science/paper/ROOI4P2E"},"agent_actions":{"view_html":"https://pith.science/pith/ROOI4P2EUYHAAWVMNPVPUUDKQC","download_json":"https://pith.science/pith/ROOI4P2EUYHAAWVMNPVPUUDKQC.json","view_paper":"https://pith.science/paper/ROOI4P2E","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.17709&json=true","fetch_graph":"https://pith.science/api/pith-number/ROOI4P2EUYHAAWVMNPVPUUDKQC/graph.json","fetch_events":"https://pith.science/api/pith-number/ROOI4P2EUYHAAWVMNPVPUUDKQC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ROOI4P2EUYHAAWVMNPVPUUDKQC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ROOI4P2EUYHAAWVMNPVPUUDKQC/action/storage_attestation","attest_author":"https://pith.science/pith/ROOI4P2EUYHAAWVMNPVPUUDKQC/action/author_attestation","sign_citation":"https://pith.science/pith/ROOI4P2EUYHAAWVMNPVPUUDKQC/action/citation_signature","submit_replication":"https://pith.science/pith/ROOI4P2EUYHAAWVMNPVPUUDKQC/action/replication_record"}},"created_at":"2026-07-05T11:33:25.030924+00:00","updated_at":"2026-07-05T11:33:25.030924+00:00"}