{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:SZFWBOATNRZ6UJ5YU544FLMNET","short_pith_number":"pith:SZFWBOAT","canonical_record":{"source":{"id":"2605.13421","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ME","submitted_at":"2026-05-13T12:16:04Z","cross_cats_sorted":[],"title_canon_sha256":"f885738edb92e46abd9a2b6170399ee325ca88d5506bac25f255507728c4df07","abstract_canon_sha256":"ff52b2eab230b2afdef90eef9cffde8b363aa0fd9c8151e3ad7af745130b60b3"},"schema_version":"1.0"},"canonical_sha256":"964b60b8136c73ea27b8a779c2ad8d24fc81dcb047ab1d34823f91a2d8f06224","source":{"kind":"arxiv","id":"2605.13421","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.13421","created_at":"2026-05-18T02:44:47Z"},{"alias_kind":"arxiv_version","alias_value":"2605.13421v1","created_at":"2026-05-18T02:44:47Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.13421","created_at":"2026-05-18T02:44:47Z"},{"alias_kind":"pith_short_12","alias_value":"SZFWBOATNRZ6","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"SZFWBOATNRZ6UJ5Y","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"SZFWBOAT","created_at":"2026-05-18T12:33:37Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:SZFWBOATNRZ6UJ5YU544FLMNET","target":"record","payload":{"canonical_record":{"source":{"id":"2605.13421","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ME","submitted_at":"2026-05-13T12:16:04Z","cross_cats_sorted":[],"title_canon_sha256":"f885738edb92e46abd9a2b6170399ee325ca88d5506bac25f255507728c4df07","abstract_canon_sha256":"ff52b2eab230b2afdef90eef9cffde8b363aa0fd9c8151e3ad7af745130b60b3"},"schema_version":"1.0"},"canonical_sha256":"964b60b8136c73ea27b8a779c2ad8d24fc81dcb047ab1d34823f91a2d8f06224","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T02:44:47.333569Z","signature_b64":"wGi6CC7yaaSmeMU+nVespqkpD/HEFxXI0q8kJFwjO5HFIXLnd2Izdwzel2gpZvZvhxFzr3kE2YXnsTiXx4pZBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"964b60b8136c73ea27b8a779c2ad8d24fc81dcb047ab1d34823f91a2d8f06224","last_reissued_at":"2026-05-18T02:44:47.333101Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T02:44:47.333101Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.13421","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":"2026-05-18T02:44:47Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"R5t6q3o7H4aUO+Y+82dUB1W0T7x0hF4iEr1Z7kyEaID2CrWaXpXG2Ert2SYfBtlyOpxH/aiRZR7D7iJUlUoWAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-29T21:08:50.340733Z"},"content_sha256":"cabe4db72e68b1414e1135824afe20a1f0b9b38a691dd43bd37b2397bdb2527a","schema_version":"1.0","event_id":"sha256:cabe4db72e68b1414e1135824afe20a1f0b9b38a691dd43bd37b2397bdb2527a"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:SZFWBOATNRZ6UJ5YU544FLMNET","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Combining pre-trained models via localized model averaging","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Modeling averaging weights as functions of covariates yields asymptotically optimal in-sample and out-of-sample risks when combining pre-trained models.","cross_cats":[],"primary_cat":"stat.ME","authors_text":"Baihua He, Yuhong Yang, Ziwen Gao","submitted_at":"2026-05-13T12:16:04Z","abstract_excerpt":"Many pre-trained models (PTMs) are available in modern applications. Because different PTMs are often trained on different datasets, their performances can vary substantially for different new tasks, and the ranking of the candidates may depend heavily on the input. Motivated by this, we propose a localized model averaging method with weights modeled as functions of the covariates, making it substantially more versatile than existing model averaging methods. This formulation allows the model averaging procedure to adaptively capture the varying relative advantages of different PTMs across hete"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"We establish the asymptotic optimality of the proposed method for both in-sample and out-of-sample risks, as well as the consistency of the estimated weights.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The local weights can be flexibly learned as functions of covariates under a general loss that accommodates broad prediction tasks, with data conditions allowing consistent estimation.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Localized model averaging with covariate-dependent weights achieves asymptotic optimality and weight consistency for combining pre-trained models under a general loss framework.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Modeling averaging weights as functions of covariates yields asymptotically optimal in-sample and out-of-sample risks when combining pre-trained models.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"bb06912bf5a25caff77c2985e7767c1bee1b0eb1ccbe1fdc325a40884a5ac50c"},"source":{"id":"2605.13421","kind":"arxiv","version":1},"verdict":{"id":"f81283c4-ab63-4e62-a758-ff967a67c7cf","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-14T17:53:50.959691Z","strongest_claim":"We establish the asymptotic optimality of the proposed method for both in-sample and out-of-sample risks, as well as the consistency of the estimated weights.","one_line_summary":"Localized model averaging with covariate-dependent weights achieves asymptotic optimality and weight consistency for combining pre-trained models under a general loss framework.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The local weights can be flexibly learned as functions of covariates under a general loss that accommodates broad prediction tasks, with data conditions allowing consistent estimation.","pith_extraction_headline":"Modeling averaging weights as functions of covariates yields asymptotically optimal in-sample and out-of-sample risks when combining pre-trained models."},"references":{"count":202,"sample":[{"doi":"","year":2000,"title":"The Annals of Statistics , volume=","work_id":"4bbfc97a-de26-49b3-93e3-9ae76af76cbe","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2004,"title":"Combining forecasting procedures:","work_id":"f0f6b37e-5036-41d8-b741-a239ba372dce","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2013,"title":"Journal of Econometrics , volume=","work_id":"d5f7081c-01fe-4f1d-b7ca-a0cea5c215bf","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2022,"title":"Journal of the American Statistical Association , volume=","work_id":"5e2206ad-f041-4618-af1d-2bbdd666c51c","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2017,"title":"Economics Letters , year=2017, volume=","work_id":"da69008e-fe96-459e-9a4c-d1eead510182","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":202,"snapshot_sha256":"83914a00b78e89c3961ca49d2d61c841447f5f1b82fe1321adc49da8b228c0a8","internal_anchors":5},"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":"f81283c4-ab63-4e62-a758-ff967a67c7cf"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-18T02:44:47Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"HEtC+LD5H/lNaSndhNQFiLp9XOaG2Om7+DKDiIRG9XB7rq4uqIgH73Ruzj4thv6GQzcRsbZY0k6JBzRE3sfUDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-29T21:08:50.341239Z"},"content_sha256":"99cfbb6f3d68a2be5fd73679dc13fc4c31f0d7f857c8ab023ba2d4ea3b1cb7a4","schema_version":"1.0","event_id":"sha256:99cfbb6f3d68a2be5fd73679dc13fc4c31f0d7f857c8ab023ba2d4ea3b1cb7a4"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/SZFWBOATNRZ6UJ5YU544FLMNET/bundle.json","state_url":"https://pith.science/pith/SZFWBOATNRZ6UJ5YU544FLMNET/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/SZFWBOATNRZ6UJ5YU544FLMNET/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":"2026-05-29T21:08:50Z","links":{"resolver":"https://pith.science/pith/SZFWBOATNRZ6UJ5YU544FLMNET","bundle":"https://pith.science/pith/SZFWBOATNRZ6UJ5YU544FLMNET/bundle.json","state":"https://pith.science/pith/SZFWBOATNRZ6UJ5YU544FLMNET/state.json","well_known_bundle":"https://pith.science/.well-known/pith/SZFWBOATNRZ6UJ5YU544FLMNET/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:SZFWBOATNRZ6UJ5YU544FLMNET","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":"ff52b2eab230b2afdef90eef9cffde8b363aa0fd9c8151e3ad7af745130b60b3","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ME","submitted_at":"2026-05-13T12:16:04Z","title_canon_sha256":"f885738edb92e46abd9a2b6170399ee325ca88d5506bac25f255507728c4df07"},"schema_version":"1.0","source":{"id":"2605.13421","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.13421","created_at":"2026-05-18T02:44:47Z"},{"alias_kind":"arxiv_version","alias_value":"2605.13421v1","created_at":"2026-05-18T02:44:47Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.13421","created_at":"2026-05-18T02:44:47Z"},{"alias_kind":"pith_short_12","alias_value":"SZFWBOATNRZ6","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"SZFWBOATNRZ6UJ5Y","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"SZFWBOAT","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:99cfbb6f3d68a2be5fd73679dc13fc4c31f0d7f857c8ab023ba2d4ea3b1cb7a4","target":"graph","created_at":"2026-05-18T02:44:47Z","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":"We establish the asymptotic optimality of the proposed method for both in-sample and out-of-sample risks, as well as the consistency of the estimated weights."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"The local weights can be flexibly learned as functions of covariates under a general loss that accommodates broad prediction tasks, with data conditions allowing consistent estimation."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Localized model averaging with covariate-dependent weights achieves asymptotic optimality and weight consistency for combining pre-trained models under a general loss framework."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Modeling averaging weights as functions of covariates yields asymptotically optimal in-sample and out-of-sample risks when combining pre-trained models."}],"snapshot_sha256":"bb06912bf5a25caff77c2985e7767c1bee1b0eb1ccbe1fdc325a40884a5ac50c"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"paper":{"abstract_excerpt":"Many pre-trained models (PTMs) are available in modern applications. Because different PTMs are often trained on different datasets, their performances can vary substantially for different new tasks, and the ranking of the candidates may depend heavily on the input. Motivated by this, we propose a localized model averaging method with weights modeled as functions of the covariates, making it substantially more versatile than existing model averaging methods. This formulation allows the model averaging procedure to adaptively capture the varying relative advantages of different PTMs across hete","authors_text":"Baihua He, Yuhong Yang, Ziwen Gao","cross_cats":[],"headline":"Modeling averaging weights as functions of covariates yields asymptotically optimal in-sample and out-of-sample risks when combining pre-trained models.","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ME","submitted_at":"2026-05-13T12:16:04Z","title":"Combining pre-trained models via localized model averaging"},"references":{"count":202,"internal_anchors":5,"resolved_work":202,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"The Annals of Statistics , volume=","work_id":"4bbfc97a-de26-49b3-93e3-9ae76af76cbe","year":2000},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"Combining forecasting procedures:","work_id":"f0f6b37e-5036-41d8-b741-a239ba372dce","year":2004},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Journal of Econometrics , volume=","work_id":"d5f7081c-01fe-4f1d-b7ca-a0cea5c215bf","year":2013},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"Journal of the American Statistical Association , volume=","work_id":"5e2206ad-f041-4618-af1d-2bbdd666c51c","year":2022},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"Economics Letters , year=2017, volume=","work_id":"da69008e-fe96-459e-9a4c-d1eead510182","year":2017}],"snapshot_sha256":"83914a00b78e89c3961ca49d2d61c841447f5f1b82fe1321adc49da8b228c0a8"},"source":{"id":"2605.13421","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-14T17:53:50.959691Z","id":"f81283c4-ab63-4e62-a758-ff967a67c7cf","model_set":{"reader":"grok-4.3"},"one_line_summary":"Localized model averaging with covariate-dependent weights achieves asymptotic optimality and weight consistency for combining pre-trained models under a general loss framework.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Modeling averaging weights as functions of covariates yields asymptotically optimal in-sample and out-of-sample risks when combining pre-trained models.","strongest_claim":"We establish the asymptotic optimality of the proposed method for both in-sample and out-of-sample risks, as well as the consistency of the estimated weights.","weakest_assumption":"The local weights can be flexibly learned as functions of covariates under a general loss that accommodates broad prediction tasks, with data conditions allowing consistent estimation."}},"verdict_id":"f81283c4-ab63-4e62-a758-ff967a67c7cf"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:cabe4db72e68b1414e1135824afe20a1f0b9b38a691dd43bd37b2397bdb2527a","target":"record","created_at":"2026-05-18T02:44:47Z","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":"ff52b2eab230b2afdef90eef9cffde8b363aa0fd9c8151e3ad7af745130b60b3","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ME","submitted_at":"2026-05-13T12:16:04Z","title_canon_sha256":"f885738edb92e46abd9a2b6170399ee325ca88d5506bac25f255507728c4df07"},"schema_version":"1.0","source":{"id":"2605.13421","kind":"arxiv","version":1}},"canonical_sha256":"964b60b8136c73ea27b8a779c2ad8d24fc81dcb047ab1d34823f91a2d8f06224","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"964b60b8136c73ea27b8a779c2ad8d24fc81dcb047ab1d34823f91a2d8f06224","first_computed_at":"2026-05-18T02:44:47.333101Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-18T02:44:47.333101Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"wGi6CC7yaaSmeMU+nVespqkpD/HEFxXI0q8kJFwjO5HFIXLnd2Izdwzel2gpZvZvhxFzr3kE2YXnsTiXx4pZBA==","signature_status":"signed_v1","signed_at":"2026-05-18T02:44:47.333569Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.13421","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:cabe4db72e68b1414e1135824afe20a1f0b9b38a691dd43bd37b2397bdb2527a","sha256:99cfbb6f3d68a2be5fd73679dc13fc4c31f0d7f857c8ab023ba2d4ea3b1cb7a4"],"state_sha256":"3202d1cf794cfcd3b10813ebb2d953cd0a05fe33ee32a42d6730b1deb44d7608"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"crJYhbDpRcWLtP8x1s+I+YLGdY0zfYGcn1XlWSG22yuwfUa8CcDvEuFdgh1G/vD/PuBNDGk5+C6SCv3dKQ/oBQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-29T21:08:50.343664Z","bundle_sha256":"e6ca6e16dc5d5748a42458de589ec55705fd2445968b51747d68d93cfdadc14d"}}