{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:ZWACN2K2EGMR3KWGUNPUSNVWGR","short_pith_number":"pith:ZWACN2K2","schema_version":"1.0","canonical_sha256":"cd8026e95a21991daac6a35f4936b6347b1de694098ec5a65777331e34f2f87b","source":{"kind":"arxiv","id":"2508.15015","version":1},"attestation_state":"computed","paper":{"title":"Fragment-Wise Interpretability in Graph Neural Networks via Molecule Decomposition and Contribution Analysis","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Bartosz Zieli\\'nski, Sebastian Musia{\\l}, Tomasz Danel","submitted_at":"2025-08-20T19:15:53Z","abstract_excerpt":"Graph neural networks have demonstrated remarkable success in predicting molecular properties by leveraging the rich structural information encoded in molecular graphs. However, their black-box nature reduces interpretability, which limits trust in their predictions for important applications such as drug discovery and materials design. Furthermore, existing explanation techniques often fail to reliably quantify the contribution of individual atoms or substructures due to the entangled message-passing dynamics. We introduce SEAL (Substructure Explanation via Attribution Learning), a new interp"},"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":"2508.15015","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-08-20T19:15:53Z","cross_cats_sorted":[],"title_canon_sha256":"a1fd87a5c13961c4c34dfd362cc51cc3cd22b90622a9ac8475c848fb70c8c3ba","abstract_canon_sha256":"705adc50ab0eb3dab9db29332c453d0e4dd23e4522b6ee28fcd8adf4986f82b4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:57:05.103272Z","signature_b64":"JlYk+Nf8McslQa71DBPgRyI7J6N9EGcs5V6uODXpMQOkzIkkDVOzxxLR4frYgsT4NDBnPCm9UT7FRQqh1gRbCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"cd8026e95a21991daac6a35f4936b6347b1de694098ec5a65777331e34f2f87b","last_reissued_at":"2026-07-05T11:57:05.102839Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:57:05.102839Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Fragment-Wise Interpretability in Graph Neural Networks via Molecule Decomposition and Contribution Analysis","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Bartosz Zieli\\'nski, Sebastian Musia{\\l}, Tomasz Danel","submitted_at":"2025-08-20T19:15:53Z","abstract_excerpt":"Graph neural networks have demonstrated remarkable success in predicting molecular properties by leveraging the rich structural information encoded in molecular graphs. However, their black-box nature reduces interpretability, which limits trust in their predictions for important applications such as drug discovery and materials design. Furthermore, existing explanation techniques often fail to reliably quantify the contribution of individual atoms or substructures due to the entangled message-passing dynamics. We introduce SEAL (Substructure Explanation via Attribution Learning), a new interp"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.15015","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/2508.15015/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":"2508.15015","created_at":"2026-07-05T11:57:05.102895+00:00"},{"alias_kind":"arxiv_version","alias_value":"2508.15015v1","created_at":"2026-07-05T11:57:05.102895+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.15015","created_at":"2026-07-05T11:57:05.102895+00:00"},{"alias_kind":"pith_short_12","alias_value":"ZWACN2K2EGMR","created_at":"2026-07-05T11:57:05.102895+00:00"},{"alias_kind":"pith_short_16","alias_value":"ZWACN2K2EGMR3KWG","created_at":"2026-07-05T11:57:05.102895+00:00"},{"alias_kind":"pith_short_8","alias_value":"ZWACN2K2","created_at":"2026-07-05T11:57:05.102895+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.10230","citing_title":"FORGE: Fragment-Oriented Ranking and Generation for Context-Aware Molecular Optimization","ref_index":21,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ZWACN2K2EGMR3KWGUNPUSNVWGR","json":"https://pith.science/pith/ZWACN2K2EGMR3KWGUNPUSNVWGR.json","graph_json":"https://pith.science/api/pith-number/ZWACN2K2EGMR3KWGUNPUSNVWGR/graph.json","events_json":"https://pith.science/api/pith-number/ZWACN2K2EGMR3KWGUNPUSNVWGR/events.json","paper":"https://pith.science/paper/ZWACN2K2"},"agent_actions":{"view_html":"https://pith.science/pith/ZWACN2K2EGMR3KWGUNPUSNVWGR","download_json":"https://pith.science/pith/ZWACN2K2EGMR3KWGUNPUSNVWGR.json","view_paper":"https://pith.science/paper/ZWACN2K2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2508.15015&json=true","fetch_graph":"https://pith.science/api/pith-number/ZWACN2K2EGMR3KWGUNPUSNVWGR/graph.json","fetch_events":"https://pith.science/api/pith-number/ZWACN2K2EGMR3KWGUNPUSNVWGR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ZWACN2K2EGMR3KWGUNPUSNVWGR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ZWACN2K2EGMR3KWGUNPUSNVWGR/action/storage_attestation","attest_author":"https://pith.science/pith/ZWACN2K2EGMR3KWGUNPUSNVWGR/action/author_attestation","sign_citation":"https://pith.science/pith/ZWACN2K2EGMR3KWGUNPUSNVWGR/action/citation_signature","submit_replication":"https://pith.science/pith/ZWACN2K2EGMR3KWGUNPUSNVWGR/action/replication_record"}},"created_at":"2026-07-05T11:57:05.102895+00:00","updated_at":"2026-07-05T11:57:05.102895+00:00"}