{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:S33IQNG7QDOHW2LLBSGDN5QP55","short_pith_number":"pith:S33IQNG7","canonical_record":{"source":{"id":"2605.12619","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"q-bio.NC","submitted_at":"2026-05-12T18:06:57Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"cec122090485572f7c7ab3726a6b2e88e3caafcc91b8e8ec357ee9049723a9b8","abstract_canon_sha256":"4be21ebe381454fdd136f4761217f600a7baf07ac78891834a7ec661f3af9a9d"},"schema_version":"1.0"},"canonical_sha256":"96f68834df80dc7b696b0c8c36f60fef41c8a314b084daef66335217d5d5d8e5","source":{"kind":"arxiv","id":"2605.12619","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.12619","created_at":"2026-05-18T03:10:00Z"},{"alias_kind":"arxiv_version","alias_value":"2605.12619v1","created_at":"2026-05-18T03:10:00Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.12619","created_at":"2026-05-18T03:10:00Z"},{"alias_kind":"pith_short_12","alias_value":"S33IQNG7QDOH","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"S33IQNG7QDOHW2LL","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"S33IQNG7","created_at":"2026-05-18T12:33:37Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:S33IQNG7QDOHW2LLBSGDN5QP55","target":"record","payload":{"canonical_record":{"source":{"id":"2605.12619","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"q-bio.NC","submitted_at":"2026-05-12T18:06:57Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"cec122090485572f7c7ab3726a6b2e88e3caafcc91b8e8ec357ee9049723a9b8","abstract_canon_sha256":"4be21ebe381454fdd136f4761217f600a7baf07ac78891834a7ec661f3af9a9d"},"schema_version":"1.0"},"canonical_sha256":"96f68834df80dc7b696b0c8c36f60fef41c8a314b084daef66335217d5d5d8e5","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T03:10:00.461303Z","signature_b64":"WGrbKIAJuAQbY8qxk4DOg3osBJZsBvVO8BrzNWdGSteGZwCY7XxFc2ilsajszYFi2Ui9JtyN5ZwND/lrcXxrDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"96f68834df80dc7b696b0c8c36f60fef41c8a314b084daef66335217d5d5d8e5","last_reissued_at":"2026-05-18T03:10:00.460670Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T03:10:00.460670Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.12619","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-18T03:10:00Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"CaMAqAJCRB7n0DwD/nzlVu01MvB+bjFhtkALqGArkBYwFyqPHzvbloNulp9KvyNEpgUmLoNanZj0R3seFZT/Dw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-21T17:37:20.018735Z"},"content_sha256":"2227f5e00150c4701d4c33895586d055238034900ace8171dd4d02ef36722895","schema_version":"1.0","event_id":"sha256:2227f5e00150c4701d4c33895586d055238034900ace8171dd4d02ef36722895"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:S33IQNG7QDOHW2LLBSGDN5QP55","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Human face perception reflects inverse-generative and naturalistic discriminative objectives","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Human face perception aligns best with neural networks trained via inverse rendering or natural-image classification.","cross_cats":["cs.CV"],"primary_cat":"q-bio.NC","authors_text":"Heiko H. Sch\\\"utt, Kamila Maria Jozwik, Katherine R. Storrs, Nikolaus Kriegeskorte, Tal Golan, Wenxuan Guo","submitted_at":"2026-05-12T18:06:57Z","abstract_excerpt":"The perceptual representations supporting our ability to recognize faces remain a computational mystery. Deep neural networks offer mechanistic hypotheses for human face perception, but theoretically distinct models often make indistinguishable representational predictions for randomly sampled faces. To expose diagnostic differences among these hypotheses, we compared six neural network models sharing an architecture but trained on distinct tasks, using face pairs optimized to elicit contrasting model predictions (\"controversial\" pairs) alongside randomly sampled pairs. We tested model predict"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Models prioritizing high-level, invariant structures (trained via inverse rendering, face identification, or object classification) most robustly matched human judgments. Furthermore, models trained on natural images typically outperformed synthetic-trained counterparts.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the controversial face pairs, optimized to elicit contrasting model predictions, accurately isolate the perceptual dimensions relevant to humans without introducing optimization artifacts or biases in the stimulus generation process.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Human face perception aligns with neural networks trained on inverse-generative and naturalistic discriminative tasks, as these best predict human dissimilarity judgments on controversial and random face pairs.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Human face perception aligns best with neural networks trained via inverse rendering or natural-image classification.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"0ed9434f295ebfce5e213a24207dd446dc3308433c4c2c53ead99c03ec487103"},"source":{"id":"2605.12619","kind":"arxiv","version":1},"verdict":{"id":"4ad48618-4668-4909-8409-2e8081f1b307","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-14T20:25:50.303906Z","strongest_claim":"Models prioritizing high-level, invariant structures (trained via inverse rendering, face identification, or object classification) most robustly matched human judgments. Furthermore, models trained on natural images typically outperformed synthetic-trained counterparts.","one_line_summary":"Human face perception aligns with neural networks trained on inverse-generative and naturalistic discriminative tasks, as these best predict human dissimilarity judgments on controversial and random face pairs.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the controversial face pairs, optimized to elicit contrasting model predictions, accurately isolate the perceptual dimensions relevant to humans without introducing optimization artifacts or biases in the stimulus generation process.","pith_extraction_headline":"Human face perception aligns best with neural networks trained via inverse rendering or natural-image classification."},"references":{"count":96,"sample":[{"doi":"","year":2021,"title":"O’Toole, A. J. & Castillo, C. D. Face Recognition by Humans and Machines: Three Fundamental AdvancesfromDeepLearning.Annualreviewofvisionscience7,543–570(2021). https://doi.org/10.1146/ annurev-vision","work_id":"a3b7d91e-bb10-4737-baf2-aca437d5edc0","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.1162/jocn_a_02040","year":2023,"title":"van Dyck, L. E. & Gruber, W. R. Modeling biological face recognition with deep convolutional neural networks.JournalofCognitiveNeuroscience35,1521–1537(2023). https://doi.org/10.1162/jocn_a_02040","work_id":"6574b23a-9254-47c9-a18f-5b1c81d79b3d","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.1126/sciadv.aax5979","year":2020,"title":"Yildirim, I., Belledonne, M., Freiwald, W. & Tenenbaum, J. Efficient inverse graphics in biological face processing.Science Advances6, eaax5979 (2020). https://doi.org/10.1126/sciadv.aax5979","work_id":"7705d154-1fee-499a-9bfe-87430e278751","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2021,"title":"https://doi.org/10.1016/ j.patter.2021.100348","work_id":"b68e5285-d550-4980-bab4-56ce1f6fdf3a","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.1016/j.cognition","year":2021,"title":"Blauch, N. M., Behrmann, M. & Plaut, D. C. Computational insights into human perceptual expertise for familiarandunfamiliarfacerecognition.Cognition208,104341(2021). https://doi.org/10.1016/j.cognitio","work_id":"9c02ee38-353b-4a12-ad9c-e1b666306a99","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":96,"snapshot_sha256":"c754c0e70ffb9f7d7f6b1a98c6ac684a200890bc9c30e9aa0ad103f8c0ccebf2","internal_anchors":3},"formal_canon":{"evidence_count":2,"snapshot_sha256":"d489c4c3b64b23f1008c9b49d3b8ea79778e8286079852b9a11e2c9ab71089fd"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":"4ad48618-4668-4909-8409-2e8081f1b307"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-18T03:10:00Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"yJa7cS3oD6yp5U7qmp2apmavNBfI5IlkuATUq+frnmo5nqFwefrFxLRea9nmI/X2KSUn7++bHLEkltIJIybnDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-21T17:37:20.019815Z"},"content_sha256":"e86a2dcdca308902b279169c1faff3551b43eb165b1213f95fa5d26e289c37da","schema_version":"1.0","event_id":"sha256:e86a2dcdca308902b279169c1faff3551b43eb165b1213f95fa5d26e289c37da"},{"event_type":"integrity_finding","subject_pith_number":"pith:2026:S33IQNG7QDOHW2LLBSGDN5QP55","target":"integrity","payload":{"note":"DOI in the printed bibliography is fragmented by whitespace or line breaks. A longer candidate (10.1016/j.cub.2021.04.014.21) was visible in the surrounding text but could not be confirmed against doi.org as printed.","snippet":"Chang, L., Egger, B., Vetter, T. & Tsao, D. Y. Explaining face representation in the primate brain using different computational models.Current Biology31, 2785–2795.e4 (2021). https://doi.org/10.1016/j.cub. 2021.04.014. 21","arxiv_id":"2605.12619","detector":"doi_compliance","evidence":{"ref_index":27,"verdict_class":"incontrovertible","resolved_title":null,"printed_excerpt":"Chang, L., Egger, B., Vetter, T. & Tsao, D. Y. Explaining face representation in the primate brain using different computational models.Current Biology31, 2785–2795.e4 (2021). https://doi.org/10.1016/j.cub. 2021.04.014. 21","reconstructed_doi":"10.1016/j.cub.2021.04.014.21"},"severity":"advisory","ref_index":27,"audited_at":"2026-05-19T07:16:02.609100Z","event_type":"pith.integrity.v1","detected_doi":"10.1016/j.cub.2021.04.014.21","detector_url":"https://pith.science/pith-integrity-protocol#doi_compliance","external_url":null,"finding_type":"recoverable_identifier","evidence_hash":"a80d69b53b50d30386624000a70abf0c0d1a79579dc3bae7b28ce25ad7a3e7b8","paper_version":1,"verdict_class":"incontrovertible","resolved_title":null,"detector_version":"1.0.0","detected_arxiv_id":null,"integrity_event_id":284,"payload_sha256":"a34bcae8a10382bf7e2561cb2b94b52bc0687cc60a51873d2c1ac34efe9d0ef8","signature_b64":"sjFM2VpVtAN5/noB93pQ8FpFlLBebMWX/6wRR8JtUciWrvvXTiftpe7zA9mcAEg44uTT+rKzth7bk+Hy6f9PBQ==","signing_key_id":"pith-v1-2026-05"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-19T07:16:54Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"5AJjVOSQK21uYi4WV65HZTJaSkGI7wlALJXzjdLW1yLWJK5Ij1L5z9g6pDYy/LAOL/InDvR3yxxF441c4OsiDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-21T17:37:20.021117Z"},"content_sha256":"addd0643cd64adabac6f5dd652015438284c241175a4a781e53373b4a9ff8aeb","schema_version":"1.0","event_id":"sha256:addd0643cd64adabac6f5dd652015438284c241175a4a781e53373b4a9ff8aeb"},{"event_type":"integrity_finding","subject_pith_number":"pith:2026:S33IQNG7QDOHW2LLBSGDN5QP55","target":"integrity","payload":{"note":"Identifier '10.1109/cvpr.2017' is syntactically valid but the DOI registry (doi.org) returned 404, and Crossref / OpenAlex / internal corpus also have no record. The cited work could not be located through any authoritative source.","snippet":"Liu, W.et al.SphereFace: Deep Hypersphere Embedding for Face Recognition.2017 IEEE Conference on ComputerVisionandPatternRecognition(CVPR),6738–6746(2017). https://doi.org/10.1109/CVPR.2017","arxiv_id":"2605.12619","detector":"doi_compliance","evidence":{"doi":"10.1109/cvpr.2017","arxiv_id":null,"ref_index":16,"raw_excerpt":"Liu, W.et al.SphereFace: Deep Hypersphere Embedding for Face Recognition.2017 IEEE Conference on ComputerVisionandPatternRecognition(CVPR),6738–6746(2017). https://doi.org/10.1109/CVPR.2017","verdict_class":"cross_source","checked_sources":["crossref_by_doi","openalex_by_doi","doi_org_head"]},"severity":"critical","ref_index":16,"audited_at":"2026-05-19T07:16:02.609100Z","event_type":"pith.integrity.v1","detected_doi":"10.1109/cvpr.2017","detector_url":"https://pith.science/pith-integrity-protocol#doi_compliance","external_url":null,"finding_type":"unresolvable_identifier","evidence_hash":"72988a90dfc9b3fba93b78ad27751711d9fc14e4a27b0345d42962f3746a5cf3","paper_version":1,"verdict_class":"cross_source","resolved_title":null,"detector_version":"1.0.0","detected_arxiv_id":null,"integrity_event_id":283,"payload_sha256":"3a1a1835d8dfe8e2c39a625b926e84712f397f0695d2e72f4558a4b034bf556a","signature_b64":"mN6NnUpOVntISGVbQ+JulBvG4zBowKlE3E1S4OkU25TEF/t+tBL2O+cl0VNsYcaE94Lgr8SFygylaZuU6mYKAg==","signing_key_id":"pith-v1-2026-05"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-19T07:16:54Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"XLNyBR7B/oFR2weq666ezr3LuLtbja+VoZ3lv0LKseJ++r5KyZ98g6mWl+pRlAVe+bNXBn/5Yz5SQqubASSvAw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-21T17:37:20.021469Z"},"content_sha256":"4311c142121afde638303f59bc5d8e8f2bac650dfd7d217260661c87d54ebd22","schema_version":"1.0","event_id":"sha256:4311c142121afde638303f59bc5d8e8f2bac650dfd7d217260661c87d54ebd22"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/S33IQNG7QDOHW2LLBSGDN5QP55/bundle.json","state_url":"https://pith.science/pith/S33IQNG7QDOHW2LLBSGDN5QP55/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/S33IQNG7QDOHW2LLBSGDN5QP55/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":"2026-05-21T17:37:20Z","links":{"resolver":"https://pith.science/pith/S33IQNG7QDOHW2LLBSGDN5QP55","bundle":"https://pith.science/pith/S33IQNG7QDOHW2LLBSGDN5QP55/bundle.json","state":"https://pith.science/pith/S33IQNG7QDOHW2LLBSGDN5QP55/state.json","well_known_bundle":"https://pith.science/.well-known/pith/S33IQNG7QDOHW2LLBSGDN5QP55/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:S33IQNG7QDOHW2LLBSGDN5QP55","merge_version":"pith-open-graph-merge-v1","event_count":4,"valid_event_count":4,"invalid_event_count":0,"equivocation_count":1,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"4be21ebe381454fdd136f4761217f600a7baf07ac78891834a7ec661f3af9a9d","cross_cats_sorted":["cs.CV"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"q-bio.NC","submitted_at":"2026-05-12T18:06:57Z","title_canon_sha256":"cec122090485572f7c7ab3726a6b2e88e3caafcc91b8e8ec357ee9049723a9b8"},"schema_version":"1.0","source":{"id":"2605.12619","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.12619","created_at":"2026-05-18T03:10:00Z"},{"alias_kind":"arxiv_version","alias_value":"2605.12619v1","created_at":"2026-05-18T03:10:00Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.12619","created_at":"2026-05-18T03:10:00Z"},{"alias_kind":"pith_short_12","alias_value":"S33IQNG7QDOH","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"S33IQNG7QDOHW2LL","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"S33IQNG7","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:e86a2dcdca308902b279169c1faff3551b43eb165b1213f95fa5d26e289c37da","target":"graph","created_at":"2026-05-18T03:10:00Z","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":"Models prioritizing high-level, invariant structures (trained via inverse rendering, face identification, or object classification) most robustly matched human judgments. Furthermore, models trained on natural images typically outperformed synthetic-trained counterparts."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That the controversial face pairs, optimized to elicit contrasting model predictions, accurately isolate the perceptual dimensions relevant to humans without introducing optimization artifacts or biases in the stimulus generation process."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Human face perception aligns with neural networks trained on inverse-generative and naturalistic discriminative tasks, as these best predict human dissimilarity judgments on controversial and random face pairs."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Human face perception aligns best with neural networks trained via inverse rendering or natural-image classification."}],"snapshot_sha256":"0ed9434f295ebfce5e213a24207dd446dc3308433c4c2c53ead99c03ec487103"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"d489c4c3b64b23f1008c9b49d3b8ea79778e8286079852b9a11e2c9ab71089fd"},"paper":{"abstract_excerpt":"The perceptual representations supporting our ability to recognize faces remain a computational mystery. Deep neural networks offer mechanistic hypotheses for human face perception, but theoretically distinct models often make indistinguishable representational predictions for randomly sampled faces. To expose diagnostic differences among these hypotheses, we compared six neural network models sharing an architecture but trained on distinct tasks, using face pairs optimized to elicit contrasting model predictions (\"controversial\" pairs) alongside randomly sampled pairs. We tested model predict","authors_text":"Heiko H. Sch\\\"utt, Kamila Maria Jozwik, Katherine R. Storrs, Nikolaus Kriegeskorte, Tal Golan, Wenxuan Guo","cross_cats":["cs.CV"],"headline":"Human face perception aligns best with neural networks trained via inverse rendering or natural-image classification.","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"q-bio.NC","submitted_at":"2026-05-12T18:06:57Z","title":"Human face perception reflects inverse-generative and naturalistic discriminative objectives"},"references":{"count":96,"internal_anchors":3,"resolved_work":96,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"O’Toole, A. J. & Castillo, C. D. Face Recognition by Humans and Machines: Three Fundamental AdvancesfromDeepLearning.Annualreviewofvisionscience7,543–570(2021). https://doi.org/10.1146/ annurev-vision","work_id":"a3b7d91e-bb10-4737-baf2-aca437d5edc0","year":2021},{"cited_arxiv_id":"","doi":"10.1162/jocn_a_02040","is_internal_anchor":false,"ref_index":2,"title":"van Dyck, L. E. & Gruber, W. R. Modeling biological face recognition with deep convolutional neural networks.JournalofCognitiveNeuroscience35,1521–1537(2023). https://doi.org/10.1162/jocn_a_02040","work_id":"6574b23a-9254-47c9-a18f-5b1c81d79b3d","year":2023},{"cited_arxiv_id":"","doi":"10.1126/sciadv.aax5979","is_internal_anchor":false,"ref_index":3,"title":"Yildirim, I., Belledonne, M., Freiwald, W. & Tenenbaum, J. Efficient inverse graphics in biological face processing.Science Advances6, eaax5979 (2020). https://doi.org/10.1126/sciadv.aax5979","work_id":"7705d154-1fee-499a-9bfe-87430e278751","year":2020},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"https://doi.org/10.1016/ j.patter.2021.100348","work_id":"b68e5285-d550-4980-bab4-56ce1f6fdf3a","year":2021},{"cited_arxiv_id":"","doi":"10.1016/j.cognition","is_internal_anchor":false,"ref_index":5,"title":"Blauch, N. M., Behrmann, M. & Plaut, D. C. Computational insights into human perceptual expertise for familiarandunfamiliarfacerecognition.Cognition208,104341(2021). https://doi.org/10.1016/j.cognitio","work_id":"9c02ee38-353b-4a12-ad9c-e1b666306a99","year":2021}],"snapshot_sha256":"c754c0e70ffb9f7d7f6b1a98c6ac684a200890bc9c30e9aa0ad103f8c0ccebf2"},"source":{"id":"2605.12619","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-14T20:25:50.303906Z","id":"4ad48618-4668-4909-8409-2e8081f1b307","model_set":{"reader":"grok-4.3"},"one_line_summary":"Human face perception aligns with neural networks trained on inverse-generative and naturalistic discriminative tasks, as these best predict human dissimilarity judgments on controversial and random face pairs.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Human face perception aligns best with neural networks trained via inverse rendering or natural-image classification.","strongest_claim":"Models prioritizing high-level, invariant structures (trained via inverse rendering, face identification, or object classification) most robustly matched human judgments. Furthermore, models trained on natural images typically outperformed synthetic-trained counterparts.","weakest_assumption":"That the controversial face pairs, optimized to elicit contrasting model predictions, accurately isolate the perceptual dimensions relevant to humans without introducing optimization artifacts or biases in the stimulus generation process."}},"verdict_id":"4ad48618-4668-4909-8409-2e8081f1b307"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:2227f5e00150c4701d4c33895586d055238034900ace8171dd4d02ef36722895","target":"record","created_at":"2026-05-18T03:10:00Z","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":"4be21ebe381454fdd136f4761217f600a7baf07ac78891834a7ec661f3af9a9d","cross_cats_sorted":["cs.CV"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"q-bio.NC","submitted_at":"2026-05-12T18:06:57Z","title_canon_sha256":"cec122090485572f7c7ab3726a6b2e88e3caafcc91b8e8ec357ee9049723a9b8"},"schema_version":"1.0","source":{"id":"2605.12619","kind":"arxiv","version":1}},"canonical_sha256":"96f68834df80dc7b696b0c8c36f60fef41c8a314b084daef66335217d5d5d8e5","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"96f68834df80dc7b696b0c8c36f60fef41c8a314b084daef66335217d5d5d8e5","first_computed_at":"2026-05-18T03:10:00.460670Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-18T03:10:00.460670Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"WGrbKIAJuAQbY8qxk4DOg3osBJZsBvVO8BrzNWdGSteGZwCY7XxFc2ilsajszYFi2Ui9JtyN5ZwND/lrcXxrDw==","signature_status":"signed_v1","signed_at":"2026-05-18T03:10:00.461303Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.12619","source_kind":"arxiv","source_version":1}}},"equivocations":[{"signer_id":"pith.science","event_type":"integrity_finding","target":"integrity","event_ids":["sha256:4311c142121afde638303f59bc5d8e8f2bac650dfd7d217260661c87d54ebd22","sha256:addd0643cd64adabac6f5dd652015438284c241175a4a781e53373b4a9ff8aeb"]}],"invalid_events":[],"applied_event_ids":["sha256:2227f5e00150c4701d4c33895586d055238034900ace8171dd4d02ef36722895","sha256:e86a2dcdca308902b279169c1faff3551b43eb165b1213f95fa5d26e289c37da"],"state_sha256":"7b7247a953e552cbbb557d9ad420d212559ff49186f25d0e74ef6c693aa203e4"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"5ScR1JxLawwku9kThGwOXr+6Y2GSd2Is/zOX9Bz9vDW8SUHk/xaGAtFtDrk0mtl0uVu8UJYpvkRdYNPvlTprAA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-21T17:37:20.025063Z","bundle_sha256":"215587f03fa2d9bf51e08083ce12b4c98274a50ce1ec04690e889caf662b6d27"}}