{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:B7ZBUQTU2772JHIZAIU5YNYUIK","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"b76083594955f239bcce2f0bca83fc88e2a335982334b286cc0309cd16e36be7","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-09T10:14:46Z","title_canon_sha256":"f1a378f7b19b36c88511ece79a7cf78dd59a24d8b92302c8f1a341e73a48ddf5"},"schema_version":"1.0","source":{"id":"2605.08858","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.08858","created_at":"2026-05-21T01:05:20Z"},{"alias_kind":"arxiv_version","alias_value":"2605.08858v2","created_at":"2026-05-21T01:05:20Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.08858","created_at":"2026-05-21T01:05:20Z"},{"alias_kind":"pith_short_12","alias_value":"B7ZBUQTU2772","created_at":"2026-05-21T01:05:20Z"},{"alias_kind":"pith_short_16","alias_value":"B7ZBUQTU2772JHIZ","created_at":"2026-05-21T01:05:20Z"},{"alias_kind":"pith_short_8","alias_value":"B7ZBUQTU","created_at":"2026-05-21T01:05:20Z"}],"graph_snapshots":[{"event_id":"sha256:0d8c97ad07016efa5d8b98c9240a79eb82c2ce25e8e342f3b1f94f817f0f9792","target":"graph","created_at":"2026-05-21T01:05:20Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"a novel framework that leverages generative models to synthesize pure, high-fidelity prototypes directly from the frozen model's weights, completely eliminating the dependency on any external data."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That generative models can reliably produce prototypes whose visual and semantic properties match those that would have been selected from real data, and that these synthetic prototypes yield faithful explanations for the original model's decisions."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"ProDG uses generative models to create data-free prototypes from model weights for post-hoc explainability of image predictions."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"ProDG generates high-fidelity visual prototypes for explaining neural network decisions using only the model's weights and no real data."}],"snapshot_sha256":"11d9a7a2801a274cb8f7ceebba0422842c9e7c19261fa7d074d623e2da10e945"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[{"findings_count":0,"name":"claim_evidence","ran_at":"2026-05-20T08:42:02.002463Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"ai_meta_artifact","ran_at":"2026-05-19T22:33:37.527474Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_title_agreement","ran_at":"2026-05-19T14:01:21.677684Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_compliance","ran_at":"2026-05-19T10:42:21.814802Z","status":"completed","version":"1.0.0"}],"endpoint":"/pith/2605.08858/integrity.json","findings":[],"snapshot_sha256":"d2b23cb9a0bf9cb54d41eb71fd5daca23597073ea5af75e2d3e5852cf15bf065","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Ante-hoc interpretability methods based on prototypes provide highly accurate explanations by utilizing the intuitive \"this looks like that\" reasoning paradigm. On the other hand, post-hoc models can explain predictions for a single image without relying on an underlying dataset or requiring costly neural network retraining. Recent approaches successfully solve the retraining problem for prototype-based networks. However, they still face a fundamental limitation: they require access to a subset of data (e.g., a test or validation set) to search for and extract the visual prototypes. In this pa","authors_text":"Jacek Tabor, {\\L}ukasz Struski, Magdalena Tr\\k{e}dowicz, Piotr Borycki, Przemys{\\l}aw Spurek","cross_cats":[],"headline":"ProDG generates high-fidelity visual prototypes for explaining neural network decisions using only the model's weights and no real data.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-09T10:14:46Z","title":"ProDG: Prototypes for Data-Free Generative Post-Hoc Explainability"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.08858","kind":"arxiv","version":2},"verdict":{"created_at":"2026-05-12T01:19:33.162363Z","id":"180c68e7-5e9a-44e0-bf34-37be5354b09d","model_set":{"reader":"grok-4.3"},"one_line_summary":"ProDG uses generative models to create data-free prototypes from model weights for post-hoc explainability of image predictions.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"ProDG generates high-fidelity visual prototypes for explaining neural network decisions using only the model's weights and no real data.","strongest_claim":"a novel framework that leverages generative models to synthesize pure, high-fidelity prototypes directly from the frozen model's weights, completely eliminating the dependency on any external data.","weakest_assumption":"That generative models can reliably produce prototypes whose visual and semantic properties match those that would have been selected from real data, and that these synthetic prototypes yield faithful explanations for the original model's decisions."}},"verdict_id":"180c68e7-5e9a-44e0-bf34-37be5354b09d"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:946b256c18e6368dd480cfe30cab4902434b904965652143f3e8fa34af6883e1","target":"record","created_at":"2026-05-21T01:05:20Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"b76083594955f239bcce2f0bca83fc88e2a335982334b286cc0309cd16e36be7","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-09T10:14:46Z","title_canon_sha256":"f1a378f7b19b36c88511ece79a7cf78dd59a24d8b92302c8f1a341e73a48ddf5"},"schema_version":"1.0","source":{"id":"2605.08858","kind":"arxiv","version":2}},"canonical_sha256":"0ff21a4274d7ffa49d190229dc371442a7f8110f04f28f5d682b29d3a2a1e4ea","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"0ff21a4274d7ffa49d190229dc371442a7f8110f04f28f5d682b29d3a2a1e4ea","first_computed_at":"2026-05-21T01:05:20.705874Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-21T01:05:20.705874Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"56/bialDfp2uV0+zbAx38R2Mvlz4pm7YM803VfWIOFB6Bxluez4JIyoiLeTIq+lRJrhg5XXalHsuHf6tgTn/BQ==","signature_status":"signed_v1","signed_at":"2026-05-21T01:05:20.706423Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.08858","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:946b256c18e6368dd480cfe30cab4902434b904965652143f3e8fa34af6883e1","sha256:0d8c97ad07016efa5d8b98c9240a79eb82c2ce25e8e342f3b1f94f817f0f9792"],"state_sha256":"46eabdb7978c5d69644178de8cdb64ce08d33945e6a4fb7486e78d349bf9d70a"}