{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:DCPCCDBZHZTCR2IFNXWM3VGXSW","short_pith_number":"pith:DCPCCDBZ","schema_version":"1.0","canonical_sha256":"189e210c393e6628e9056deccdd4d795a691c7b12a9dc5f4043313d8a030e5f8","source":{"kind":"arxiv","id":"2109.02040","version":1},"attestation_state":"computed","paper":{"title":"Data Efficient Masked Language Modeling for Vision and Language","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV","cs.LG"],"primary_cat":"cs.CL","authors_text":"Gabriel Stanovsky, Michael Elhadad, Roy Schwartz, Yonatan Bitton","submitted_at":"2021-09-05T11:27:53Z","abstract_excerpt":"Masked language modeling (MLM) is one of the key sub-tasks in vision-language pretraining. In the cross-modal setting, tokens in the sentence are masked at random, and the model predicts the masked tokens given the image and the text. In this paper, we observe several key disadvantages of MLM in this setting. First, as captions tend to be short, in a third of the sentences no token is sampled. Second, the majority of masked tokens are stop-words and punctuation, leading to under-utilization of the image. We investigate a range of alternative masking strategies specific to the cross-modal setti"},"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":"2109.02040","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2021-09-05T11:27:53Z","cross_cats_sorted":["cs.CV","cs.LG"],"title_canon_sha256":"562819232c7eec647499eb4e4a8f30728c4bc9dafc1e71a20aa972e3853d3a95","abstract_canon_sha256":"5c1c2f98792650f5f6a4ddb78ac1390fc56f9f24f2858d92a2a26310e93ed3b4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:11:37.868952Z","signature_b64":"lCyRKz4YQMQrmVvffUn5AI2VzVAUhUkBe/1FfsjmGDd7aUJSY7349QPW02WZfeF+0jgG6kCW0Tis1d5A9VoVBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"189e210c393e6628e9056deccdd4d795a691c7b12a9dc5f4043313d8a030e5f8","last_reissued_at":"2026-07-05T03:11:37.868537Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:11:37.868537Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Data Efficient Masked Language Modeling for Vision and Language","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV","cs.LG"],"primary_cat":"cs.CL","authors_text":"Gabriel Stanovsky, Michael Elhadad, Roy Schwartz, Yonatan Bitton","submitted_at":"2021-09-05T11:27:53Z","abstract_excerpt":"Masked language modeling (MLM) is one of the key sub-tasks in vision-language pretraining. In the cross-modal setting, tokens in the sentence are masked at random, and the model predicts the masked tokens given the image and the text. In this paper, we observe several key disadvantages of MLM in this setting. First, as captions tend to be short, in a third of the sentences no token is sampled. Second, the majority of masked tokens are stop-words and punctuation, leading to under-utilization of the image. We investigate a range of alternative masking strategies specific to the cross-modal setti"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2109.02040","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/2109.02040/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":"2109.02040","created_at":"2026-07-05T03:11:37.868596+00:00"},{"alias_kind":"arxiv_version","alias_value":"2109.02040v1","created_at":"2026-07-05T03:11:37.868596+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2109.02040","created_at":"2026-07-05T03:11:37.868596+00:00"},{"alias_kind":"pith_short_12","alias_value":"DCPCCDBZHZTC","created_at":"2026-07-05T03:11:37.868596+00:00"},{"alias_kind":"pith_short_16","alias_value":"DCPCCDBZHZTCR2IF","created_at":"2026-07-05T03:11:37.868596+00:00"},{"alias_kind":"pith_short_8","alias_value":"DCPCCDBZ","created_at":"2026-07-05T03:11:37.868596+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/DCPCCDBZHZTCR2IFNXWM3VGXSW","json":"https://pith.science/pith/DCPCCDBZHZTCR2IFNXWM3VGXSW.json","graph_json":"https://pith.science/api/pith-number/DCPCCDBZHZTCR2IFNXWM3VGXSW/graph.json","events_json":"https://pith.science/api/pith-number/DCPCCDBZHZTCR2IFNXWM3VGXSW/events.json","paper":"https://pith.science/paper/DCPCCDBZ"},"agent_actions":{"view_html":"https://pith.science/pith/DCPCCDBZHZTCR2IFNXWM3VGXSW","download_json":"https://pith.science/pith/DCPCCDBZHZTCR2IFNXWM3VGXSW.json","view_paper":"https://pith.science/paper/DCPCCDBZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2109.02040&json=true","fetch_graph":"https://pith.science/api/pith-number/DCPCCDBZHZTCR2IFNXWM3VGXSW/graph.json","fetch_events":"https://pith.science/api/pith-number/DCPCCDBZHZTCR2IFNXWM3VGXSW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DCPCCDBZHZTCR2IFNXWM3VGXSW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DCPCCDBZHZTCR2IFNXWM3VGXSW/action/storage_attestation","attest_author":"https://pith.science/pith/DCPCCDBZHZTCR2IFNXWM3VGXSW/action/author_attestation","sign_citation":"https://pith.science/pith/DCPCCDBZHZTCR2IFNXWM3VGXSW/action/citation_signature","submit_replication":"https://pith.science/pith/DCPCCDBZHZTCR2IFNXWM3VGXSW/action/replication_record"}},"created_at":"2026-07-05T03:11:37.868596+00:00","updated_at":"2026-07-05T03:11:37.868596+00:00"}