{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:U3GW37ZZIIKJ3AUB7F5VWKVQUF","short_pith_number":"pith:U3GW37ZZ","canonical_record":{"source":{"id":"2605.13936","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-13T16:20:33Z","cross_cats_sorted":["cs.AI","cs.DC"],"title_canon_sha256":"f470e2b794b8517051d523037c819e9b256eecd9bad4a1b287a4b68cf6e9e43d","abstract_canon_sha256":"f13a6e858b74d002e504efd33fe91a11c25c23a83b24c06f1f480a533c191e22"},"schema_version":"1.0"},"canonical_sha256":"a6cd6dff3942149d8281f97b5b2ab0a170e35424d1e0ac2c9298edf57c7a5bef","source":{"kind":"arxiv","id":"2605.13936","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.13936","created_at":"2026-05-17T23:39:13Z"},{"alias_kind":"arxiv_version","alias_value":"2605.13936v1","created_at":"2026-05-17T23:39:13Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.13936","created_at":"2026-05-17T23:39:13Z"},{"alias_kind":"pith_short_12","alias_value":"U3GW37ZZIIKJ","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"U3GW37ZZIIKJ3AUB","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"U3GW37ZZ","created_at":"2026-05-18T12:33:37Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:U3GW37ZZIIKJ3AUB7F5VWKVQUF","target":"record","payload":{"canonical_record":{"source":{"id":"2605.13936","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-13T16:20:33Z","cross_cats_sorted":["cs.AI","cs.DC"],"title_canon_sha256":"f470e2b794b8517051d523037c819e9b256eecd9bad4a1b287a4b68cf6e9e43d","abstract_canon_sha256":"f13a6e858b74d002e504efd33fe91a11c25c23a83b24c06f1f480a533c191e22"},"schema_version":"1.0"},"canonical_sha256":"a6cd6dff3942149d8281f97b5b2ab0a170e35424d1e0ac2c9298edf57c7a5bef","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:39:13.918275Z","signature_b64":"GnH+FN0Z2kKkk2Af2ahm3obKFHgEzQtSH/aWF7KXDtp9STkEsVHVcQOgLQxoLy7w9aB+phM8+gsJWKxopQzSDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a6cd6dff3942149d8281f97b5b2ab0a170e35424d1e0ac2c9298edf57c7a5bef","last_reissued_at":"2026-05-17T23:39:13.917614Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:39:13.917614Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.13936","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":"2026-05-17T23:39:13Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"iGnEbB43gYhN9dCkipssD/b1dWE31Q0amCdmgSTu0C71r0aLypM0i23fGYGlW8acD7f1qvXCwf68+LfJdqLcBg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-03T21:34:46.149158Z"},"content_sha256":"b05f2241b7f80be1ddf85ef69133e6d02b23eb3b83df877c54f0a15b628df175","schema_version":"1.0","event_id":"sha256:b05f2241b7f80be1ddf85ef69133e6d02b23eb3b83df877c54f0a15b628df175"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:U3GW37ZZIIKJ3AUB7F5VWKVQUF","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Towards the Next Frontier of LLMs, Training on Private Data: A Cross-Domain Benchmark for Federated Fine-Tuning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Federated fine-tuning lets LLMs adapt to private institutional data in healthcare and finance while matching centralized training performance.","cross_cats":["cs.AI","cs.DC"],"primary_cat":"cs.LG","authors_text":"Daniel M. Jimenez-Gutierrez, Enrique Zuazua, Georgios Kellaris, Joaquin Del Rio, Oleksii Sliusarenko, Xabi Uribe-Etxebarria","submitted_at":"2026-05-13T16:20:33Z","abstract_excerpt":"The recent success of large language models (LLMs) has been largely driven by vast public datasets. However, the next frontier for LLM development lies beyond public data. Much of the world's most valuable information is private, especially in highly regulated sectors such as healthcare and finance, where data include patient histories or customer communications. Unlocking this data could represent a major leap forward, enabling LLMs with deeper domain expertise and stronger real-world utility. Yet, these data cannot be shared because they are distributed across institutions and constrained by"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Our results show that federated fine-tuning performs close to centralized training and outperforms isolated single-institution learning.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The chosen non-IID partitions and four datasets sufficiently capture real institutional heterogeneity in population, modalities, and label distributions.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Federated PEFT on LLMs across healthcare and finance datasets performs close to centralized training and beats isolated local training under non-IID conditions.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Federated fine-tuning lets LLMs adapt to private institutional data in healthcare and finance while matching centralized training performance.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"1394a265750363b87fdc598446a235925b89fc083af675616def9264a4f80ee2"},"source":{"id":"2605.13936","kind":"arxiv","version":1},"verdict":{"id":"98a6cd53-8663-4cc8-a9e2-0612b9f2ace3","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-15T04:52:35.149189Z","strongest_claim":"Our results show that federated fine-tuning performs close to centralized training and outperforms isolated single-institution learning.","one_line_summary":"Federated PEFT on LLMs across healthcare and finance datasets performs close to centralized training and beats isolated local training under non-IID conditions.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The chosen non-IID partitions and four datasets sufficiently capture real institutional heterogeneity in population, modalities, and label distributions.","pith_extraction_headline":"Federated fine-tuning lets LLMs adapt to private institutional data in healthcare and finance while matching centralized training performance."},"references":{"count":38,"sample":[{"doi":"","year":2022,"title":"Training language models to follow instructions with human feedback.Advances in neural information processing systems, 35:27730--27744, 2022","work_id":"80b434d0-1e95-4c46-94b8-38c063076a85","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"Large language models in the clinic: a comprehensive benchmark.arXiv preprint arXiv:2405.00716, 2024","work_id":"fb746cd2-a9e0-4e8c-90d8-18d738d737c8","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"Open finllm leaderboard: Towards financial ai readiness, 2025","work_id":"865cea9d-7093-4391-b520-79fd986a237b","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2022,"title":"Lora: Low-rank adaptation of large language models.Iclr, 1(2):3","work_id":"80a84b35-042e-4846-80ca-177dbf9b1b1b","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"Qlora: Efficient finetuning of quantized llms.Advances in neural information processing systems, 36:10088--10115, 2023","work_id":"9f1e9c55-a6ee-4ab2-9db5-eb59a81ca82a","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":38,"snapshot_sha256":"8bce03288f2b0fd868133bd65e8acc71aeff5b02480ac49d2a8824a7a6507c13","internal_anchors":2},"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":"98a6cd53-8663-4cc8-a9e2-0612b9f2ace3"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:39:13Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"7xTXK40jnFYR8/z/mU/HhzSrO/CNqqwjizSiWrPYv5SUorsbazZdkBlLtBZfSipoyayXkSqNBzSw4Bf6lWSLDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-03T21:34:46.149705Z"},"content_sha256":"cc47a16bffbe32e8290db4b1efdb79e71a8bdaa71aa4235b92ee2a872a1b2e0e","schema_version":"1.0","event_id":"sha256:cc47a16bffbe32e8290db4b1efdb79e71a8bdaa71aa4235b92ee2a872a1b2e0e"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/U3GW37ZZIIKJ3AUB7F5VWKVQUF/bundle.json","state_url":"https://pith.science/pith/U3GW37ZZIIKJ3AUB7F5VWKVQUF/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/U3GW37ZZIIKJ3AUB7F5VWKVQUF/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":"2026-06-03T21:34:46Z","links":{"resolver":"https://pith.science/pith/U3GW37ZZIIKJ3AUB7F5VWKVQUF","bundle":"https://pith.science/pith/U3GW37ZZIIKJ3AUB7F5VWKVQUF/bundle.json","state":"https://pith.science/pith/U3GW37ZZIIKJ3AUB7F5VWKVQUF/state.json","well_known_bundle":"https://pith.science/.well-known/pith/U3GW37ZZIIKJ3AUB7F5VWKVQUF/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:U3GW37ZZIIKJ3AUB7F5VWKVQUF","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":"f13a6e858b74d002e504efd33fe91a11c25c23a83b24c06f1f480a533c191e22","cross_cats_sorted":["cs.AI","cs.DC"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-13T16:20:33Z","title_canon_sha256":"f470e2b794b8517051d523037c819e9b256eecd9bad4a1b287a4b68cf6e9e43d"},"schema_version":"1.0","source":{"id":"2605.13936","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.13936","created_at":"2026-05-17T23:39:13Z"},{"alias_kind":"arxiv_version","alias_value":"2605.13936v1","created_at":"2026-05-17T23:39:13Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.13936","created_at":"2026-05-17T23:39:13Z"},{"alias_kind":"pith_short_12","alias_value":"U3GW37ZZIIKJ","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"U3GW37ZZIIKJ3AUB","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"U3GW37ZZ","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:cc47a16bffbe32e8290db4b1efdb79e71a8bdaa71aa4235b92ee2a872a1b2e0e","target":"graph","created_at":"2026-05-17T23:39:13Z","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":"Our results show that federated fine-tuning performs close to centralized training and outperforms isolated single-institution learning."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"The chosen non-IID partitions and four datasets sufficiently capture real institutional heterogeneity in population, modalities, and label distributions."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Federated PEFT on LLMs across healthcare and finance datasets performs close to centralized training and beats isolated local training under non-IID conditions."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Federated fine-tuning lets LLMs adapt to private institutional data in healthcare and finance while matching centralized training performance."}],"snapshot_sha256":"1394a265750363b87fdc598446a235925b89fc083af675616def9264a4f80ee2"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"paper":{"abstract_excerpt":"The recent success of large language models (LLMs) has been largely driven by vast public datasets. However, the next frontier for LLM development lies beyond public data. Much of the world's most valuable information is private, especially in highly regulated sectors such as healthcare and finance, where data include patient histories or customer communications. Unlocking this data could represent a major leap forward, enabling LLMs with deeper domain expertise and stronger real-world utility. Yet, these data cannot be shared because they are distributed across institutions and constrained by","authors_text":"Daniel M. Jimenez-Gutierrez, Enrique Zuazua, Georgios Kellaris, Joaquin Del Rio, Oleksii Sliusarenko, Xabi Uribe-Etxebarria","cross_cats":["cs.AI","cs.DC"],"headline":"Federated fine-tuning lets LLMs adapt to private institutional data in healthcare and finance while matching centralized training performance.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-13T16:20:33Z","title":"Towards the Next Frontier of LLMs, Training on Private Data: A Cross-Domain Benchmark for Federated Fine-Tuning"},"references":{"count":38,"internal_anchors":2,"resolved_work":38,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"Training language models to follow instructions with human feedback.Advances in neural information processing systems, 35:27730--27744, 2022","work_id":"80b434d0-1e95-4c46-94b8-38c063076a85","year":2022},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"Large language models in the clinic: a comprehensive benchmark.arXiv preprint arXiv:2405.00716, 2024","work_id":"fb746cd2-a9e0-4e8c-90d8-18d738d737c8","year":2024},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Open finllm leaderboard: Towards financial ai readiness, 2025","work_id":"865cea9d-7093-4391-b520-79fd986a237b","year":2025},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"Lora: Low-rank adaptation of large language models.Iclr, 1(2):3","work_id":"80a84b35-042e-4846-80ca-177dbf9b1b1b","year":2022},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"Qlora: Efficient finetuning of quantized llms.Advances in neural information processing systems, 36:10088--10115, 2023","work_id":"9f1e9c55-a6ee-4ab2-9db5-eb59a81ca82a","year":2023}],"snapshot_sha256":"8bce03288f2b0fd868133bd65e8acc71aeff5b02480ac49d2a8824a7a6507c13"},"source":{"id":"2605.13936","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-15T04:52:35.149189Z","id":"98a6cd53-8663-4cc8-a9e2-0612b9f2ace3","model_set":{"reader":"grok-4.3"},"one_line_summary":"Federated PEFT on LLMs across healthcare and finance datasets performs close to centralized training and beats isolated local training under non-IID conditions.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Federated fine-tuning lets LLMs adapt to private institutional data in healthcare and finance while matching centralized training performance.","strongest_claim":"Our results show that federated fine-tuning performs close to centralized training and outperforms isolated single-institution learning.","weakest_assumption":"The chosen non-IID partitions and four datasets sufficiently capture real institutional heterogeneity in population, modalities, and label distributions."}},"verdict_id":"98a6cd53-8663-4cc8-a9e2-0612b9f2ace3"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:b05f2241b7f80be1ddf85ef69133e6d02b23eb3b83df877c54f0a15b628df175","target":"record","created_at":"2026-05-17T23:39:13Z","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":"f13a6e858b74d002e504efd33fe91a11c25c23a83b24c06f1f480a533c191e22","cross_cats_sorted":["cs.AI","cs.DC"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-13T16:20:33Z","title_canon_sha256":"f470e2b794b8517051d523037c819e9b256eecd9bad4a1b287a4b68cf6e9e43d"},"schema_version":"1.0","source":{"id":"2605.13936","kind":"arxiv","version":1}},"canonical_sha256":"a6cd6dff3942149d8281f97b5b2ab0a170e35424d1e0ac2c9298edf57c7a5bef","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"a6cd6dff3942149d8281f97b5b2ab0a170e35424d1e0ac2c9298edf57c7a5bef","first_computed_at":"2026-05-17T23:39:13.917614Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-17T23:39:13.917614Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"GnH+FN0Z2kKkk2Af2ahm3obKFHgEzQtSH/aWF7KXDtp9STkEsVHVcQOgLQxoLy7w9aB+phM8+gsJWKxopQzSDA==","signature_status":"signed_v1","signed_at":"2026-05-17T23:39:13.918275Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.13936","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:b05f2241b7f80be1ddf85ef69133e6d02b23eb3b83df877c54f0a15b628df175","sha256:cc47a16bffbe32e8290db4b1efdb79e71a8bdaa71aa4235b92ee2a872a1b2e0e"],"state_sha256":"a743611ade0bb6c378de1793143c4fb4bbfcda3a9f668d09aee946d77b359aac"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"7en9VEXxQ4EYXhm9EL2zzgFWwWGBRrkj0P2XR91Rv62cZ0QW7XRldA4UJeu/EuEzrspUsUMd1ImiyeeP4vO0Ag==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-03T21:34:46.152171Z","bundle_sha256":"2d6da96cfc1eb0c41089013289021993486af288dae2f1a6c6dd2ef90fabe3ef"}}