{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:QGDOPDLHJEHDB3PUOHPE7YLJKA","short_pith_number":"pith:QGDOPDLH","schema_version":"1.0","canonical_sha256":"8186e78d67490e30edf471de4fe1695005f7769d795cdcd78b4c9e14f2e2fa7b","source":{"kind":"arxiv","id":"2303.15747","version":4},"attestation_state":"computed","paper":{"title":"TabRet: Pre-training Transformer-based Tabular Models for Unseen Columns","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Kenta Oono, Kohei Hayashi, Soma Onishi","submitted_at":"2023-03-28T06:03:41Z","abstract_excerpt":"We present \\emph{TabRet}, a pre-trainable Transformer-based model for tabular data. TabRet is designed to work on a downstream task that contains columns not seen in pre-training. Unlike other methods, TabRet has an extra learning step before fine-tuning called \\emph{retokenizing}, which calibrates feature embeddings based on the masked autoencoding loss. In experiments, we pre-trained TabRet with a large collection of public health surveys and fine-tuned it on classification tasks in healthcare, and TabRet achieved the best AUC performance on four datasets. In addition, an ablation study show"},"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":"2303.15747","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2023-03-28T06:03:41Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"af82cbf992eddbc8c1b6a58073dbf88ae8740333decb2aaf1066bf51c9d32939","abstract_canon_sha256":"f972365b3683b109a5d54d1bfd4a9c9dc2fab198c90e4a45955e29ae291876b2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:01:18.311813Z","signature_b64":"RUbKIor3lfBfY8z53DtpmX2iHfk91snMar4ueQB0dZhozKtrEmN9ktHhpUiYYRagAAEFonkozsTMDw2qLFnFDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8186e78d67490e30edf471de4fe1695005f7769d795cdcd78b4c9e14f2e2fa7b","last_reissued_at":"2026-07-05T06:01:18.311341Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:01:18.311341Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"TabRet: Pre-training Transformer-based Tabular Models for Unseen Columns","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Kenta Oono, Kohei Hayashi, Soma Onishi","submitted_at":"2023-03-28T06:03:41Z","abstract_excerpt":"We present \\emph{TabRet}, a pre-trainable Transformer-based model for tabular data. TabRet is designed to work on a downstream task that contains columns not seen in pre-training. Unlike other methods, TabRet has an extra learning step before fine-tuning called \\emph{retokenizing}, which calibrates feature embeddings based on the masked autoencoding loss. In experiments, we pre-trained TabRet with a large collection of public health surveys and fine-tuned it on classification tasks in healthcare, and TabRet achieved the best AUC performance on four datasets. In addition, an ablation study show"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2303.15747","kind":"arxiv","version":4},"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/2303.15747/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":"2303.15747","created_at":"2026-07-05T06:01:18.311397+00:00"},{"alias_kind":"arxiv_version","alias_value":"2303.15747v4","created_at":"2026-07-05T06:01:18.311397+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2303.15747","created_at":"2026-07-05T06:01:18.311397+00:00"},{"alias_kind":"pith_short_12","alias_value":"QGDOPDLHJEHD","created_at":"2026-07-05T06:01:18.311397+00:00"},{"alias_kind":"pith_short_16","alias_value":"QGDOPDLHJEHDB3PU","created_at":"2026-07-05T06:01:18.311397+00:00"},{"alias_kind":"pith_short_8","alias_value":"QGDOPDLH","created_at":"2026-07-05T06:01:18.311397+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2504.16109","citing_title":"Representation Learning for Tabular Data: A Comprehensive Survey","ref_index":72,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/QGDOPDLHJEHDB3PUOHPE7YLJKA","json":"https://pith.science/pith/QGDOPDLHJEHDB3PUOHPE7YLJKA.json","graph_json":"https://pith.science/api/pith-number/QGDOPDLHJEHDB3PUOHPE7YLJKA/graph.json","events_json":"https://pith.science/api/pith-number/QGDOPDLHJEHDB3PUOHPE7YLJKA/events.json","paper":"https://pith.science/paper/QGDOPDLH"},"agent_actions":{"view_html":"https://pith.science/pith/QGDOPDLHJEHDB3PUOHPE7YLJKA","download_json":"https://pith.science/pith/QGDOPDLHJEHDB3PUOHPE7YLJKA.json","view_paper":"https://pith.science/paper/QGDOPDLH","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2303.15747&json=true","fetch_graph":"https://pith.science/api/pith-number/QGDOPDLHJEHDB3PUOHPE7YLJKA/graph.json","fetch_events":"https://pith.science/api/pith-number/QGDOPDLHJEHDB3PUOHPE7YLJKA/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QGDOPDLHJEHDB3PUOHPE7YLJKA/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QGDOPDLHJEHDB3PUOHPE7YLJKA/action/storage_attestation","attest_author":"https://pith.science/pith/QGDOPDLHJEHDB3PUOHPE7YLJKA/action/author_attestation","sign_citation":"https://pith.science/pith/QGDOPDLHJEHDB3PUOHPE7YLJKA/action/citation_signature","submit_replication":"https://pith.science/pith/QGDOPDLHJEHDB3PUOHPE7YLJKA/action/replication_record"}},"created_at":"2026-07-05T06:01:18.311397+00:00","updated_at":"2026-07-05T06:01:18.311397+00:00"}