{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:CH5QOPMFGAZIZDYFUW7BAVVRSE","short_pith_number":"pith:CH5QOPMF","schema_version":"1.0","canonical_sha256":"11fb073d8530328c8f05a5be1056b191335317ae1ad2398b082357545fe657d6","source":{"kind":"arxiv","id":"2512.01667","version":3},"attestation_state":"computed","paper":{"title":"Detecting Model Misspecification in Bayesian Inverse Problems via Variational Gradient Descent","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Comparing the standard Bayesian posterior to a predictively oriented mixing distribution detects model misspecification.","cross_cats":["stat.CO"],"primary_cat":"stat.ME","authors_text":"Andrew Curtis, Chris. J. Oates, Katherine Tant, Matthew A. Fisher, Qingyang Liu, Xuebin Zhao, Zheyang Shen","submitted_at":"2025-12-01T13:37:02Z","abstract_excerpt":"Bayesian inference is optimal when the statistical model is well-specified, while outside this setting Bayesian inference can catastrophically fail; accordingly a wealth of post-Bayesian methodologies have been proposed. Predictively oriented (PrO) approaches lift the statistical model $P_\\theta$ to an (infinite) mixture model $\\int P_\\theta \\; \\mathrm{d}Q(\\theta)$ and fit this predictive distribution via minimising an entropy-regularised objective functional. In the well-specified setting one expects the mixing distribution $Q$ to concentrate around the true data-generating parameter in the l"},"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":true},"canonical_record":{"source":{"id":"2512.01667","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"stat.ME","submitted_at":"2025-12-01T13:37:02Z","cross_cats_sorted":["stat.CO"],"title_canon_sha256":"962836fbc34a32b21c54f66e709542b511b161b68dc6bcf5096b908fddae0031","abstract_canon_sha256":"36615da1b9e34a95c41aafe10ffc824f50c54de3e6971cccdd66a0987ef4201c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-08T01:03:52.143690Z","signature_b64":"+gXvwvUsB7bP/K8F28KrtDfXkUSee3N31GcRs1LYDg2mUlvuWq3Rg9Xiz+ZGoI3U3EL9PknvmH1zTRjADNHTAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"11fb073d8530328c8f05a5be1056b191335317ae1ad2398b082357545fe657d6","last_reissued_at":"2026-06-08T01:03:52.142490Z","signature_status":"signed_v1","first_computed_at":"2026-06-08T01:03:52.142490Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Detecting Model Misspecification in Bayesian Inverse Problems via Variational Gradient Descent","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Comparing the standard Bayesian posterior to a predictively oriented mixing distribution detects model misspecification.","cross_cats":["stat.CO"],"primary_cat":"stat.ME","authors_text":"Andrew Curtis, Chris. J. Oates, Katherine Tant, Matthew A. Fisher, Qingyang Liu, Xuebin Zhao, Zheyang Shen","submitted_at":"2025-12-01T13:37:02Z","abstract_excerpt":"Bayesian inference is optimal when the statistical model is well-specified, while outside this setting Bayesian inference can catastrophically fail; accordingly a wealth of post-Bayesian methodologies have been proposed. Predictively oriented (PrO) approaches lift the statistical model $P_\\theta$ to an (infinite) mixture model $\\int P_\\theta \\; \\mathrm{d}Q(\\theta)$ and fit this predictive distribution via minimising an entropy-regularised objective functional. In the well-specified setting one expects the mixing distribution $Q$ to concentrate around the true data-generating parameter in the l"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Our contribution is to demonstrate that one can empirically detect model misspecification by comparing the standard Bayesian posterior to the PrO `posterior' Q. To operationalise this, we present an efficient numerical algorithm based on variational gradient descent.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"In the well-specified setting one expects the mixing distribution Q to concentrate around the true data-generating parameter in the large data limit, while such singular concentration will typically not be observed if the model is misspecified.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Comparing the standard Bayesian posterior to a predictive-oriented mixture posterior Q fitted via variational gradient descent detects model misspecification in inverse problems.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Comparing the standard Bayesian posterior to a predictively oriented mixing distribution detects model misspecification.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"1854ac1b308944da4fdb5ac89b9a2017f8dff03d57c5b342a0723aa51de9be57"},"source":{"id":"2512.01667","kind":"arxiv","version":3},"verdict":{"id":"f99c6597-6cb3-4556-8b96-062fca4d3baf","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-17T02:54:47.786708Z","strongest_claim":"Our contribution is to demonstrate that one can empirically detect model misspecification by comparing the standard Bayesian posterior to the PrO `posterior' Q. To operationalise this, we present an efficient numerical algorithm based on variational gradient descent.","one_line_summary":"Comparing the standard Bayesian posterior to a predictive-oriented mixture posterior Q fitted via variational gradient descent detects model misspecification in inverse problems.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"In the well-specified setting one expects the mixing distribution Q to concentrate around the true data-generating parameter in the large data limit, while such singular concentration will typically not be observed if the model is misspecified.","pith_extraction_headline":"Comparing the standard Bayesian posterior to a predictively oriented mixing distribution detects model misspecification."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2512.01667/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":"fea58772bfec07b005249e249d7725de12ba9450b397a2fc834301552a1d348a"},"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":"2512.01667","created_at":"2026-06-08T01:03:52.142654+00:00"},{"alias_kind":"arxiv_version","alias_value":"2512.01667v3","created_at":"2026-06-08T01:03:52.142654+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2512.01667","created_at":"2026-06-08T01:03:52.142654+00:00"},{"alias_kind":"pith_short_12","alias_value":"CH5QOPMFGAZI","created_at":"2026-06-08T01:03:52.142654+00:00"},{"alias_kind":"pith_short_16","alias_value":"CH5QOPMFGAZIZDYF","created_at":"2026-06-08T01:03:52.142654+00:00"},{"alias_kind":"pith_short_8","alias_value":"CH5QOPMF","created_at":"2026-06-08T01:03:52.142654+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2605.00455","citing_title":"Concentration and Calibration in Predictive Bayesian Inference","ref_index":16,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/CH5QOPMFGAZIZDYFUW7BAVVRSE","json":"https://pith.science/pith/CH5QOPMFGAZIZDYFUW7BAVVRSE.json","graph_json":"https://pith.science/api/pith-number/CH5QOPMFGAZIZDYFUW7BAVVRSE/graph.json","events_json":"https://pith.science/api/pith-number/CH5QOPMFGAZIZDYFUW7BAVVRSE/events.json","paper":"https://pith.science/paper/CH5QOPMF"},"agent_actions":{"view_html":"https://pith.science/pith/CH5QOPMFGAZIZDYFUW7BAVVRSE","download_json":"https://pith.science/pith/CH5QOPMFGAZIZDYFUW7BAVVRSE.json","view_paper":"https://pith.science/paper/CH5QOPMF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2512.01667&json=true","fetch_graph":"https://pith.science/api/pith-number/CH5QOPMFGAZIZDYFUW7BAVVRSE/graph.json","fetch_events":"https://pith.science/api/pith-number/CH5QOPMFGAZIZDYFUW7BAVVRSE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CH5QOPMFGAZIZDYFUW7BAVVRSE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CH5QOPMFGAZIZDYFUW7BAVVRSE/action/storage_attestation","attest_author":"https://pith.science/pith/CH5QOPMFGAZIZDYFUW7BAVVRSE/action/author_attestation","sign_citation":"https://pith.science/pith/CH5QOPMFGAZIZDYFUW7BAVVRSE/action/citation_signature","submit_replication":"https://pith.science/pith/CH5QOPMFGAZIZDYFUW7BAVVRSE/action/replication_record"}},"created_at":"2026-06-08T01:03:52.142654+00:00","updated_at":"2026-06-08T01:03:52.142654+00:00"}