{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:VLEBBCFBEA27P5T2QZE2FIJUFT","short_pith_number":"pith:VLEBBCFB","schema_version":"1.0","canonical_sha256":"aac81088a12035f7f67a8649a2a1342ceb742c5795d9b17dc6718592f55bd40c","source":{"kind":"arxiv","id":"2512.04329","version":2},"attestation_state":"computed","paper":{"title":"A Retrieval-Augmented Generation Approach to Extracting Algorithmic Logic from Neural Networks","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.SE"],"primary_cat":"cs.CV","authors_text":"Dmitry Ignatov, Radu Timofte, Waleed Khalid","submitted_at":"2025-12-03T23:28:30Z","abstract_excerpt":"Reusing existing neural-network components is central to research efficiency, yet discovering, extracting, and validating such modules across thousands of open-source repositories remains difficult. We introduce NN-RAG, a retrieval-augmented generation system that converts large, heterogeneous PyTorch codebases into a searchable and executable library of validated neural modules. Unlike conventional code search or clone-detection tools, NN-RAG performs scope-aware dependency resolution, import-preserving reconstruction, and validator-gated promotion -- ensuring that every retrieved block is sc"},"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":"2512.04329","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-12-03T23:28:30Z","cross_cats_sorted":["cs.SE"],"title_canon_sha256":"2eef789983ae09fec792776a72d479c8fc0f0e567bcea1fc70664e807d09d13d","abstract_canon_sha256":"17af7db634772ac108a9171ea1ce112126cd138bf63833ccd45e08dd1384e429"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:02:59.720517Z","signature_b64":"6xYZIoS3jfyYC5wFgb7kvjXNikeNFBOSJTH5X+czSS+Q1Lfubx0QiQ7qocsG8DU5+lnX0XEVv5KpeduM+GwXDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"aac81088a12035f7f67a8649a2a1342ceb742c5795d9b17dc6718592f55bd40c","last_reissued_at":"2026-05-20T00:02:59.717983Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:02:59.717983Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"A Retrieval-Augmented Generation Approach to Extracting Algorithmic Logic from Neural Networks","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.SE"],"primary_cat":"cs.CV","authors_text":"Dmitry Ignatov, Radu Timofte, Waleed Khalid","submitted_at":"2025-12-03T23:28:30Z","abstract_excerpt":"Reusing existing neural-network components is central to research efficiency, yet discovering, extracting, and validating such modules across thousands of open-source repositories remains difficult. We introduce NN-RAG, a retrieval-augmented generation system that converts large, heterogeneous PyTorch codebases into a searchable and executable library of validated neural modules. Unlike conventional code search or clone-detection tools, NN-RAG performs scope-aware dependency resolution, import-preserving reconstruction, and validator-gated promotion -- ensuring that every retrieved block is sc"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2512.04329","kind":"arxiv","version":2},"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/2512.04329/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":"2512.04329","created_at":"2026-05-20T00:02:59.718136+00:00"},{"alias_kind":"arxiv_version","alias_value":"2512.04329v2","created_at":"2026-05-20T00:02:59.718136+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2512.04329","created_at":"2026-05-20T00:02:59.718136+00:00"},{"alias_kind":"pith_short_12","alias_value":"VLEBBCFBEA27","created_at":"2026-05-20T00:02:59.718136+00:00"},{"alias_kind":"pith_short_16","alias_value":"VLEBBCFBEA27P5T2","created_at":"2026-05-20T00:02:59.718136+00:00"},{"alias_kind":"pith_short_8","alias_value":"VLEBBCFB","created_at":"2026-05-20T00:02:59.718136+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":5,"sample":[{"citing_arxiv_id":"2511.07329","citing_title":"Preparation of Fractal-Inspired Computational Architectures for Advanced Large Language Model Analysis","ref_index":9,"is_internal_anchor":true},{"citing_arxiv_id":"2511.07329","citing_title":"Preparation of Fractal-Inspired Computational Architectures for Advanced Large Language Model Analysis","ref_index":9,"is_internal_anchor":true},{"citing_arxiv_id":"2512.24120","citing_title":"Enhancing LLM-Based Neural Network Generation: Few-Shot Prompting and Efficient Validation for Automated Architecture Design","ref_index":11,"is_internal_anchor":true},{"citing_arxiv_id":"2601.08517","citing_title":"Closed-Loop LLM Discovery of Non-Standard Channel Priors in Vision Models","ref_index":6,"is_internal_anchor":true},{"citing_arxiv_id":"2605.04903","citing_title":"Delta-Based Neural Architecture Search: LLM Fine-Tuning via Code Diffs","ref_index":23,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/VLEBBCFBEA27P5T2QZE2FIJUFT","json":"https://pith.science/pith/VLEBBCFBEA27P5T2QZE2FIJUFT.json","graph_json":"https://pith.science/api/pith-number/VLEBBCFBEA27P5T2QZE2FIJUFT/graph.json","events_json":"https://pith.science/api/pith-number/VLEBBCFBEA27P5T2QZE2FIJUFT/events.json","paper":"https://pith.science/paper/VLEBBCFB"},"agent_actions":{"view_html":"https://pith.science/pith/VLEBBCFBEA27P5T2QZE2FIJUFT","download_json":"https://pith.science/pith/VLEBBCFBEA27P5T2QZE2FIJUFT.json","view_paper":"https://pith.science/paper/VLEBBCFB","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2512.04329&json=true","fetch_graph":"https://pith.science/api/pith-number/VLEBBCFBEA27P5T2QZE2FIJUFT/graph.json","fetch_events":"https://pith.science/api/pith-number/VLEBBCFBEA27P5T2QZE2FIJUFT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VLEBBCFBEA27P5T2QZE2FIJUFT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VLEBBCFBEA27P5T2QZE2FIJUFT/action/storage_attestation","attest_author":"https://pith.science/pith/VLEBBCFBEA27P5T2QZE2FIJUFT/action/author_attestation","sign_citation":"https://pith.science/pith/VLEBBCFBEA27P5T2QZE2FIJUFT/action/citation_signature","submit_replication":"https://pith.science/pith/VLEBBCFBEA27P5T2QZE2FIJUFT/action/replication_record"}},"created_at":"2026-05-20T00:02:59.718136+00:00","updated_at":"2026-05-20T00:02:59.718136+00:00"}