{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:J2IXHTRNAKC43XZNB6ROY53565","short_pith_number":"pith:J2IXHTRN","canonical_record":{"source":{"id":"2505.21857","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.LG","submitted_at":"2025-05-28T01:03:28Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"90105e9e37b6dbb0d1c74267229ba30e4da7f5d394c8507803130ba7fed1069b","abstract_canon_sha256":"6ab9d46a778d4e42377a2cbc2c2d8236724a0916656187e69c489d3bcc6dd883"},"schema_version":"1.0"},"canonical_sha256":"4e9173ce2d0285cddf2d0fa2ec777df77dee18adac1dd0dbaca839f33e899e0e","source":{"kind":"arxiv","id":"2505.21857","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2505.21857","created_at":"2026-07-05T11:11:04Z"},{"alias_kind":"arxiv_version","alias_value":"2505.21857v1","created_at":"2026-07-05T11:11:04Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.21857","created_at":"2026-07-05T11:11:04Z"},{"alias_kind":"pith_short_12","alias_value":"J2IXHTRNAKC4","created_at":"2026-07-05T11:11:04Z"},{"alias_kind":"pith_short_16","alias_value":"J2IXHTRNAKC43XZN","created_at":"2026-07-05T11:11:04Z"},{"alias_kind":"pith_short_8","alias_value":"J2IXHTRN","created_at":"2026-07-05T11:11:04Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:J2IXHTRNAKC43XZNB6ROY53565","target":"record","payload":{"canonical_record":{"source":{"id":"2505.21857","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.LG","submitted_at":"2025-05-28T01:03:28Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"90105e9e37b6dbb0d1c74267229ba30e4da7f5d394c8507803130ba7fed1069b","abstract_canon_sha256":"6ab9d46a778d4e42377a2cbc2c2d8236724a0916656187e69c489d3bcc6dd883"},"schema_version":"1.0"},"canonical_sha256":"4e9173ce2d0285cddf2d0fa2ec777df77dee18adac1dd0dbaca839f33e899e0e","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:11:04.886821Z","signature_b64":"t9SOHrqJaJon4VjEMzQSh4XrWBxziH5+KKBbHUsnkSIHLF1H9/HoK0VCODasobMXYSlspxbvEO9isCx2mYK5Bw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4e9173ce2d0285cddf2d0fa2ec777df77dee18adac1dd0dbaca839f33e899e0e","last_reissued_at":"2026-07-05T11:11:04.886396Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:11:04.886396Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2505.21857","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":"2026-07-05T11:11:04Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"VuSUZzNxuoFwUs6hAjazORKLB7NIENDluwII31NVi5DRtnZ/I3SOVByTiGU0qk8JtZVanRuk+KHVsEKm4VMBAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T10:16:41.724352Z"},"content_sha256":"c8aa0916905907bfaf265073cbd2b492cc6a7f9018f9eda2a73b5da31c0bca7e","schema_version":"1.0","event_id":"sha256:c8aa0916905907bfaf265073cbd2b492cc6a7f9018f9eda2a73b5da31c0bca7e"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:J2IXHTRNAKC43XZNB6ROY53565","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Revisiting Bayesian Model Averaging in the Era of Foundation Models","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Mijung Park","submitted_at":"2025-05-28T01:03:28Z","abstract_excerpt":"We revisit the classical, full-fledged Bayesian model averaging (BMA) paradigm to ensemble pre-trained and/or lightly-finetuned foundation models to enhance the classification performance on image and text data. To make BMA tractable under foundation models, we introduce trainable linear classifiers that take frozen features from the pre-trained foundation models as inputs. The model posteriors over the linear classifiers tell us which linear heads and frozen features are better suited for a given dataset, resulting in a principled model ensembling method. Furthermore, we propose a computation"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.21857","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/2505.21857/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:11:04Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"rBBdq0UkxV82k4eActwfX+B/1qi0om1bD/mz3zwa2VBF3pDEF2xbZFuJVLrVc1q8ZxwF42Fq8WXy5WCujaFxCw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T10:16:41.724895Z"},"content_sha256":"2148b6e596412a8735e247927b39cbc4cc4a7430b9615ef1f18ebae8d037ac88","schema_version":"1.0","event_id":"sha256:2148b6e596412a8735e247927b39cbc4cc4a7430b9615ef1f18ebae8d037ac88"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/J2IXHTRNAKC43XZNB6ROY53565/bundle.json","state_url":"https://pith.science/pith/J2IXHTRNAKC43XZNB6ROY53565/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/J2IXHTRNAKC43XZNB6ROY53565/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":"2026-08-08T10:16:41Z","links":{"resolver":"https://pith.science/pith/J2IXHTRNAKC43XZNB6ROY53565","bundle":"https://pith.science/pith/J2IXHTRNAKC43XZNB6ROY53565/bundle.json","state":"https://pith.science/pith/J2IXHTRNAKC43XZNB6ROY53565/state.json","well_known_bundle":"https://pith.science/.well-known/pith/J2IXHTRNAKC43XZNB6ROY53565/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:J2IXHTRNAKC43XZNB6ROY53565","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":"6ab9d46a778d4e42377a2cbc2c2d8236724a0916656187e69c489d3bcc6dd883","cross_cats_sorted":["stat.ML"],"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.LG","submitted_at":"2025-05-28T01:03:28Z","title_canon_sha256":"90105e9e37b6dbb0d1c74267229ba30e4da7f5d394c8507803130ba7fed1069b"},"schema_version":"1.0","source":{"id":"2505.21857","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2505.21857","created_at":"2026-07-05T11:11:04Z"},{"alias_kind":"arxiv_version","alias_value":"2505.21857v1","created_at":"2026-07-05T11:11:04Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.21857","created_at":"2026-07-05T11:11:04Z"},{"alias_kind":"pith_short_12","alias_value":"J2IXHTRNAKC4","created_at":"2026-07-05T11:11:04Z"},{"alias_kind":"pith_short_16","alias_value":"J2IXHTRNAKC43XZN","created_at":"2026-07-05T11:11:04Z"},{"alias_kind":"pith_short_8","alias_value":"J2IXHTRN","created_at":"2026-07-05T11:11:04Z"}],"graph_snapshots":[{"event_id":"sha256:2148b6e596412a8735e247927b39cbc4cc4a7430b9615ef1f18ebae8d037ac88","target":"graph","created_at":"2026-07-05T11:11:04Z","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":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2505.21857/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"We revisit the classical, full-fledged Bayesian model averaging (BMA) paradigm to ensemble pre-trained and/or lightly-finetuned foundation models to enhance the classification performance on image and text data. To make BMA tractable under foundation models, we introduce trainable linear classifiers that take frozen features from the pre-trained foundation models as inputs. The model posteriors over the linear classifiers tell us which linear heads and frozen features are better suited for a given dataset, resulting in a principled model ensembling method. Furthermore, we propose a computation","authors_text":"Mijung Park","cross_cats":["stat.ML"],"headline":"","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.LG","submitted_at":"2025-05-28T01:03:28Z","title":"Revisiting Bayesian Model Averaging in the Era of Foundation Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.21857","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:c8aa0916905907bfaf265073cbd2b492cc6a7f9018f9eda2a73b5da31c0bca7e","target":"record","created_at":"2026-07-05T11:11:04Z","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":"6ab9d46a778d4e42377a2cbc2c2d8236724a0916656187e69c489d3bcc6dd883","cross_cats_sorted":["stat.ML"],"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.LG","submitted_at":"2025-05-28T01:03:28Z","title_canon_sha256":"90105e9e37b6dbb0d1c74267229ba30e4da7f5d394c8507803130ba7fed1069b"},"schema_version":"1.0","source":{"id":"2505.21857","kind":"arxiv","version":1}},"canonical_sha256":"4e9173ce2d0285cddf2d0fa2ec777df77dee18adac1dd0dbaca839f33e899e0e","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"4e9173ce2d0285cddf2d0fa2ec777df77dee18adac1dd0dbaca839f33e899e0e","first_computed_at":"2026-07-05T11:11:04.886396Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:11:04.886396Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"t9SOHrqJaJon4VjEMzQSh4XrWBxziH5+KKBbHUsnkSIHLF1H9/HoK0VCODasobMXYSlspxbvEO9isCx2mYK5Bw==","signature_status":"signed_v1","signed_at":"2026-07-05T11:11:04.886821Z","signed_message":"canonical_sha256_bytes"},"source_id":"2505.21857","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:c8aa0916905907bfaf265073cbd2b492cc6a7f9018f9eda2a73b5da31c0bca7e","sha256:2148b6e596412a8735e247927b39cbc4cc4a7430b9615ef1f18ebae8d037ac88"],"state_sha256":"7ad473109dc31f304e427945aa2f71f50cb8f6e7df59d8030891129923d033d1"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"l8L0VHaDtBCzAUCI4RDH5gBoqqcHxBUocyOG/X833wLJVmH8zxZcMq84RuhcjZXKEmRuq9bYFic7rtBhBXJWDQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-08T10:16:41.729981Z","bundle_sha256":"a3fc619a0166ac4cb268b4c732eb2e03fce931ac9d0d3b9547b1bc11f0222b12"}}