{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:CPQXFWH6ESLTI2QE7L3MLSUJ7P","short_pith_number":"pith:CPQXFWH6","schema_version":"1.0","canonical_sha256":"13e172d8fe2497346a04faf6c5ca89fbfe227e8a13853118c24bfdc2ec2c7ac5","source":{"kind":"arxiv","id":"2606.30473","version":1},"attestation_state":"computed","paper":{"title":"Field Order Should Not Matter: Permutation-Invariant Embedding Model Fine-Tuning for Structured Metadata Retrieval","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.IR","cs.LG","econ.GN","q-fin.EC"],"primary_cat":"cs.CL","authors_text":"Aivin V. Solatorio, Olivier Dupriez, Rafael Macalaba","submitted_at":"2026-06-29T15:33:04Z","abstract_excerpt":"We study retrieval over catalogs of structured metadata, where each record is a small schema whose fields answer different kinds of query. Embedding a record with a text encoder first serializes its fields into a string, which forces a choice of field order. We show this choice, usually treated as an implementation detail, silently controls retrieval quality once the encoder is fine-tuned. A standard fine-tune loses 7.4 nDCG@10 points when the index is rebuilt under a different field order, because it reads absolute position instead of the field labels. We propose permutation-invariant fine-tu"},"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":false},"canonical_record":{"source":{"id":"2606.30473","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2026-06-29T15:33:04Z","cross_cats_sorted":["cs.AI","cs.IR","cs.LG","econ.GN","q-fin.EC"],"title_canon_sha256":"37e9ab20d6ed9e501f397d0a7333a5a9559026e60d5908a36873bf29d8166576","abstract_canon_sha256":"65dfa3dad1948814d11c64f582f85a5bd5e7deccb9f7036ff11d174376dba909"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-30T02:18:16.833059Z","signature_b64":"jI5G0Sz87cLdCmV1xAN9pTTN6dlQmmRVUOFvtGRbRfKXnlNOzJZkLj0mCpBTOzy1ha2u/sLLyj7QJdb1no72AQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"13e172d8fe2497346a04faf6c5ca89fbfe227e8a13853118c24bfdc2ec2c7ac5","last_reissued_at":"2026-06-30T02:18:16.832340Z","signature_status":"signed_v1","first_computed_at":"2026-06-30T02:18:16.832340Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Field Order Should Not Matter: Permutation-Invariant Embedding Model Fine-Tuning for Structured Metadata Retrieval","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.IR","cs.LG","econ.GN","q-fin.EC"],"primary_cat":"cs.CL","authors_text":"Aivin V. Solatorio, Olivier Dupriez, Rafael Macalaba","submitted_at":"2026-06-29T15:33:04Z","abstract_excerpt":"We study retrieval over catalogs of structured metadata, where each record is a small schema whose fields answer different kinds of query. Embedding a record with a text encoder first serializes its fields into a string, which forces a choice of field order. We show this choice, usually treated as an implementation detail, silently controls retrieval quality once the encoder is fine-tuned. A standard fine-tune loses 7.4 nDCG@10 points when the index is rebuilt under a different field order, because it reads absolute position instead of the field labels. We propose permutation-invariant fine-tu"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.30473","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/2606.30473/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2606.30473","created_at":"2026-06-30T02:18:16.832447+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.30473v1","created_at":"2026-06-30T02:18:16.832447+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.30473","created_at":"2026-06-30T02:18:16.832447+00:00"},{"alias_kind":"pith_short_12","alias_value":"CPQXFWH6ESLT","created_at":"2026-06-30T02:18:16.832447+00:00"},{"alias_kind":"pith_short_16","alias_value":"CPQXFWH6ESLTI2QE","created_at":"2026-06-30T02:18:16.832447+00:00"},{"alias_kind":"pith_short_8","alias_value":"CPQXFWH6","created_at":"2026-06-30T02:18:16.832447+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/CPQXFWH6ESLTI2QE7L3MLSUJ7P","json":"https://pith.science/pith/CPQXFWH6ESLTI2QE7L3MLSUJ7P.json","graph_json":"https://pith.science/api/pith-number/CPQXFWH6ESLTI2QE7L3MLSUJ7P/graph.json","events_json":"https://pith.science/api/pith-number/CPQXFWH6ESLTI2QE7L3MLSUJ7P/events.json","paper":"https://pith.science/paper/CPQXFWH6"},"agent_actions":{"view_html":"https://pith.science/pith/CPQXFWH6ESLTI2QE7L3MLSUJ7P","download_json":"https://pith.science/pith/CPQXFWH6ESLTI2QE7L3MLSUJ7P.json","view_paper":"https://pith.science/paper/CPQXFWH6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.30473&json=true","fetch_graph":"https://pith.science/api/pith-number/CPQXFWH6ESLTI2QE7L3MLSUJ7P/graph.json","fetch_events":"https://pith.science/api/pith-number/CPQXFWH6ESLTI2QE7L3MLSUJ7P/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CPQXFWH6ESLTI2QE7L3MLSUJ7P/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CPQXFWH6ESLTI2QE7L3MLSUJ7P/action/storage_attestation","attest_author":"https://pith.science/pith/CPQXFWH6ESLTI2QE7L3MLSUJ7P/action/author_attestation","sign_citation":"https://pith.science/pith/CPQXFWH6ESLTI2QE7L3MLSUJ7P/action/citation_signature","submit_replication":"https://pith.science/pith/CPQXFWH6ESLTI2QE7L3MLSUJ7P/action/replication_record"}},"created_at":"2026-06-30T02:18:16.832447+00:00","updated_at":"2026-06-30T02:18:16.832447+00:00"}