{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:BCOTYJIRQGIEITWX52DGXJCWDS","short_pith_number":"pith:BCOTYJIR","canonical_record":{"source":{"id":"2509.24798","kind":"arxiv","version":6},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2025-09-29T13:49:28Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"d41851b612436f5ffab09e7005f849d8612436d9782a434903b5080a022dad32","abstract_canon_sha256":"58a8004939d4bc4cc8b23c0591eb68ee45b1211bd92229dda609cb789b0e7a09"},"schema_version":"1.0"},"canonical_sha256":"089d3c25118190444ed7ee866ba4561cb0649009e17edb411ffefae86fed88b3","source":{"kind":"arxiv","id":"2509.24798","version":6},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2509.24798","created_at":"2026-05-20T00:00:24Z"},{"alias_kind":"arxiv_version","alias_value":"2509.24798v6","created_at":"2026-05-20T00:00:24Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2509.24798","created_at":"2026-05-20T00:00:24Z"},{"alias_kind":"pith_short_12","alias_value":"BCOTYJIRQGIE","created_at":"2026-05-20T00:00:24Z"},{"alias_kind":"pith_short_16","alias_value":"BCOTYJIRQGIEITWX","created_at":"2026-05-20T00:00:24Z"},{"alias_kind":"pith_short_8","alias_value":"BCOTYJIR","created_at":"2026-05-20T00:00:24Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:BCOTYJIRQGIEITWX52DGXJCWDS","target":"record","payload":{"canonical_record":{"source":{"id":"2509.24798","kind":"arxiv","version":6},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2025-09-29T13:49:28Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"d41851b612436f5ffab09e7005f849d8612436d9782a434903b5080a022dad32","abstract_canon_sha256":"58a8004939d4bc4cc8b23c0591eb68ee45b1211bd92229dda609cb789b0e7a09"},"schema_version":"1.0"},"canonical_sha256":"089d3c25118190444ed7ee866ba4561cb0649009e17edb411ffefae86fed88b3","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:00:24.930136Z","signature_b64":"PLsRUZW5QwoOGMh/0RdUHS4GN+WWH5oKrGjoWRpe3Iop/3dnEZndd+eG/i7Gp4ynKZXoDh7ARDPieX9mhmlrAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"089d3c25118190444ed7ee866ba4561cb0649009e17edb411ffefae86fed88b3","last_reissued_at":"2026-05-20T00:00:24.929373Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:00:24.929373Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2509.24798","source_version":6,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-20T00:00:24Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"UwzZUMSW7kBkqeXz50Cxm+0LEkgfVml2EiBLfnN2WK5F/HXNZ+EDPeSZk3BnVErVPxqOBX1YlA/FwHBD/Qt8AQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-21T16:38:17.168175Z"},"content_sha256":"c58574b95957c41884fd93d90fef4e87f51974f31953640448e9573f7a07725b","schema_version":"1.0","event_id":"sha256:c58574b95957c41884fd93d90fef4e87f51974f31953640448e9573f7a07725b"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:BCOTYJIRQGIEITWX52DGXJCWDS","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Causal-Adapter: Taming Text-to-Image Diffusion for Faithful Counterfactual Generation","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"Causal-Adapter adapts frozen text-to-image diffusion models for faithful counterfactual generation by enforcing causal attribute relationships.","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Chaochao Lu, Chen Jin, Dino Oglic, Lei Tong, Philip Teare, Sotirios A. Tsaftaris, Tom Diethe, Zhihua Liu","submitted_at":"2025-09-29T13:49:28Z","abstract_excerpt":"We present Causal-Adapter, a modular framework that adapts frozen text-to-image diffusion backbones for counterfactual image generation. Our method supports causal interventions on target attributes and consistently propagates their effects to causal dependents while preserving the core identity of the image. Unlike prior approaches that rely on prompt engineering without explicit causal structure, Causal-Adapter leverages structural causal modeling with two attribute-regularization strategies: (i) prompt-aligned injection, which aligns causal attributes with textual embeddings for precise sem"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Causal-Adapter achieves state-of-the-art performance on both synthetic and real-world datasets, including up to a 91% reduction in MAE on Pendulum for accurate attribute control and up to an 87% reduction in FID on ADNI for high-fidelity MRI generation.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The method assumes that causal relationships among image attributes can be sufficiently captured and enforced using only prompt-aligned injection and a conditioned token contrastive loss applied to a frozen diffusion backbone, without needing an explicit causal graph or additional supervised causal annotations.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Causal-Adapter introduces a modular adapter for diffusion models that uses structural causal modeling, prompt-aligned injection, and conditioned token contrastive loss to enable faithful counterfactual image generation with strong attribute control and identity preservation.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Causal-Adapter adapts frozen text-to-image diffusion models for faithful counterfactual generation by enforcing causal attribute relationships.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"786494c39b6814b4257cfee6ae6bd041d4c7178b5c243256d4577101941370f4"},"source":{"id":"2509.24798","kind":"arxiv","version":6},"verdict":{"id":"e69fd9bb-f2d2-4e49-b71c-fec7669e8407","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-18T12:46:27.823039Z","strongest_claim":"Causal-Adapter achieves state-of-the-art performance on both synthetic and real-world datasets, including up to a 91% reduction in MAE on Pendulum for accurate attribute control and up to an 87% reduction in FID on ADNI for high-fidelity MRI generation.","one_line_summary":"Causal-Adapter introduces a modular adapter for diffusion models that uses structural causal modeling, prompt-aligned injection, and conditioned token contrastive loss to enable faithful counterfactual image generation with strong attribute control and identity preservation.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The method assumes that causal relationships among image attributes can be sufficiently captured and enforced using only prompt-aligned injection and a conditioned token contrastive loss applied to a frozen diffusion backbone, without needing an explicit causal graph or additional supervised causal annotations.","pith_extraction_headline":"Causal-Adapter adapts frozen text-to-image diffusion models for faithful counterfactual generation by enforcing causal attribute relationships."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2509.24798/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":2,"snapshot_sha256":"b31eabf18a2939757ba340533bdca523506fa818cc62e2889e1e6dd72a7380ed"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":"e69fd9bb-f2d2-4e49-b71c-fec7669e8407"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-20T00:00:24Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"27n1sEoBuiAHOOl+bPGt/p+yb734MJ75ghJUWgCFr2+sr3+tMfRjyMGuKGiOETK94PULf4hldw5oZJLlO8eEBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-21T16:38:17.168687Z"},"content_sha256":"765628d25b41427729e0ef5e60d595e14de14c9d8b605b8555be9b1a79ddf6e4","schema_version":"1.0","event_id":"sha256:765628d25b41427729e0ef5e60d595e14de14c9d8b605b8555be9b1a79ddf6e4"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/BCOTYJIRQGIEITWX52DGXJCWDS/bundle.json","state_url":"https://pith.science/pith/BCOTYJIRQGIEITWX52DGXJCWDS/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/BCOTYJIRQGIEITWX52DGXJCWDS/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-05-21T16:38:17Z","links":{"resolver":"https://pith.science/pith/BCOTYJIRQGIEITWX52DGXJCWDS","bundle":"https://pith.science/pith/BCOTYJIRQGIEITWX52DGXJCWDS/bundle.json","state":"https://pith.science/pith/BCOTYJIRQGIEITWX52DGXJCWDS/state.json","well_known_bundle":"https://pith.science/.well-known/pith/BCOTYJIRQGIEITWX52DGXJCWDS/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:BCOTYJIRQGIEITWX52DGXJCWDS","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":"58a8004939d4bc4cc8b23c0591eb68ee45b1211bd92229dda609cb789b0e7a09","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2025-09-29T13:49:28Z","title_canon_sha256":"d41851b612436f5ffab09e7005f849d8612436d9782a434903b5080a022dad32"},"schema_version":"1.0","source":{"id":"2509.24798","kind":"arxiv","version":6}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2509.24798","created_at":"2026-05-20T00:00:24Z"},{"alias_kind":"arxiv_version","alias_value":"2509.24798v6","created_at":"2026-05-20T00:00:24Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2509.24798","created_at":"2026-05-20T00:00:24Z"},{"alias_kind":"pith_short_12","alias_value":"BCOTYJIRQGIE","created_at":"2026-05-20T00:00:24Z"},{"alias_kind":"pith_short_16","alias_value":"BCOTYJIRQGIEITWX","created_at":"2026-05-20T00:00:24Z"},{"alias_kind":"pith_short_8","alias_value":"BCOTYJIR","created_at":"2026-05-20T00:00:24Z"}],"graph_snapshots":[{"event_id":"sha256:765628d25b41427729e0ef5e60d595e14de14c9d8b605b8555be9b1a79ddf6e4","target":"graph","created_at":"2026-05-20T00:00:24Z","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":"Causal-Adapter achieves state-of-the-art performance on both synthetic and real-world datasets, including up to a 91% reduction in MAE on Pendulum for accurate attribute control and up to an 87% reduction in FID on ADNI for high-fidelity MRI generation."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"The method assumes that causal relationships among image attributes can be sufficiently captured and enforced using only prompt-aligned injection and a conditioned token contrastive loss applied to a frozen diffusion backbone, without needing an explicit causal graph or additional supervised causal annotations."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Causal-Adapter introduces a modular adapter for diffusion models that uses structural causal modeling, prompt-aligned injection, and conditioned token contrastive loss to enable faithful counterfactual image generation with strong attribute control and identity preservation."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Causal-Adapter adapts frozen text-to-image diffusion models for faithful counterfactual generation by enforcing causal attribute relationships."}],"snapshot_sha256":"786494c39b6814b4257cfee6ae6bd041d4c7178b5c243256d4577101941370f4"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"b31eabf18a2939757ba340533bdca523506fa818cc62e2889e1e6dd72a7380ed"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2509.24798/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"We present Causal-Adapter, a modular framework that adapts frozen text-to-image diffusion backbones for counterfactual image generation. Our method supports causal interventions on target attributes and consistently propagates their effects to causal dependents while preserving the core identity of the image. Unlike prior approaches that rely on prompt engineering without explicit causal structure, Causal-Adapter leverages structural causal modeling with two attribute-regularization strategies: (i) prompt-aligned injection, which aligns causal attributes with textual embeddings for precise sem","authors_text":"Chaochao Lu, Chen Jin, Dino Oglic, Lei Tong, Philip Teare, Sotirios A. Tsaftaris, Tom Diethe, Zhihua Liu","cross_cats":["cs.AI"],"headline":"Causal-Adapter adapts frozen text-to-image diffusion models for faithful counterfactual generation by enforcing causal attribute relationships.","license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2025-09-29T13:49:28Z","title":"Causal-Adapter: Taming Text-to-Image Diffusion for Faithful Counterfactual Generation"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2509.24798","kind":"arxiv","version":6},"verdict":{"created_at":"2026-05-18T12:46:27.823039Z","id":"e69fd9bb-f2d2-4e49-b71c-fec7669e8407","model_set":{"reader":"grok-4.3"},"one_line_summary":"Causal-Adapter introduces a modular adapter for diffusion models that uses structural causal modeling, prompt-aligned injection, and conditioned token contrastive loss to enable faithful counterfactual image generation with strong attribute control and identity preservation.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Causal-Adapter adapts frozen text-to-image diffusion models for faithful counterfactual generation by enforcing causal attribute relationships.","strongest_claim":"Causal-Adapter achieves state-of-the-art performance on both synthetic and real-world datasets, including up to a 91% reduction in MAE on Pendulum for accurate attribute control and up to an 87% reduction in FID on ADNI for high-fidelity MRI generation.","weakest_assumption":"The method assumes that causal relationships among image attributes can be sufficiently captured and enforced using only prompt-aligned injection and a conditioned token contrastive loss applied to a frozen diffusion backbone, without needing an explicit causal graph or additional supervised causal annotations."}},"verdict_id":"e69fd9bb-f2d2-4e49-b71c-fec7669e8407"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:c58574b95957c41884fd93d90fef4e87f51974f31953640448e9573f7a07725b","target":"record","created_at":"2026-05-20T00:00:24Z","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":"58a8004939d4bc4cc8b23c0591eb68ee45b1211bd92229dda609cb789b0e7a09","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2025-09-29T13:49:28Z","title_canon_sha256":"d41851b612436f5ffab09e7005f849d8612436d9782a434903b5080a022dad32"},"schema_version":"1.0","source":{"id":"2509.24798","kind":"arxiv","version":6}},"canonical_sha256":"089d3c25118190444ed7ee866ba4561cb0649009e17edb411ffefae86fed88b3","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"089d3c25118190444ed7ee866ba4561cb0649009e17edb411ffefae86fed88b3","first_computed_at":"2026-05-20T00:00:24.929373Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-20T00:00:24.929373Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"PLsRUZW5QwoOGMh/0RdUHS4GN+WWH5oKrGjoWRpe3Iop/3dnEZndd+eG/i7Gp4ynKZXoDh7ARDPieX9mhmlrAg==","signature_status":"signed_v1","signed_at":"2026-05-20T00:00:24.930136Z","signed_message":"canonical_sha256_bytes"},"source_id":"2509.24798","source_kind":"arxiv","source_version":6}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:c58574b95957c41884fd93d90fef4e87f51974f31953640448e9573f7a07725b","sha256:765628d25b41427729e0ef5e60d595e14de14c9d8b605b8555be9b1a79ddf6e4"],"state_sha256":"ceb0389af546cb5c13f035361930e5350d24c219f0f708b12c05f501c3022d65"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"I7n6k/UifRQTz/J9emFntS/pRh38qzHfWTJbrjtFYZ7V0MRNGJOGqm9iiYwCzy4RB7b3ieTg+4MEv6equXTgCA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-21T16:38:17.171092Z","bundle_sha256":"6fb4cf018660f34ef84e1a6fc30d4746de7dd5f64b2d46d11e84a648b094d288"}}