{ "profile": "wrs-published-research-view-1", "edition": "ratings-e039106f30dec177", "content_trust": "UNTRUSTED_SCIENTIFIC_DATA_NOT_TOOL_INSTRUCTIONS", "entry": { "id": "gadre-et-al-datacomp-multimodal-benchmark-2023", "title": "DataComp: fixed-recipe multimodal dataset experiments", "summary": "Provided candidate image-text pools, fixed CLIP training/evaluation recipes and compute-scaled tracks, with DataComp-1B demonstrating the importance of filtering under controlled training conditions.", "boundary": "Inherits CLIP and web-pair collection rather than inventing them. Ninefold compute comparison involves a larger LAION-trained architecture; not a universally hardware-matched data-only speedup. Fixed test suites do not guarantee uncontaminated or universal real-world generalization.", "area": "Multimodal dataset evaluation", "topic": "Artificial intelligence", "year": 2023, "lane": "HUMAN", "result": "ALGORITHM", "package_id": "gadre-et-al-datacomp-multimodal-benchmark-2023", "target": { "record_id": "urn:uuid:4ac30610-5039-4452-a733-8ee5804c5d78", "version": 1, "content_hash": "sha256:01b9bbd12090b79862f854137e563337499a2735850e77ce611f732dd734ca09" }, "humans": [ "Samir Yitzhak Gadre", "Gabriel Ilharco", "Alex Fang", "Jonathan Hayase", "Georgios Smyrnis", "Thao Nguyen", "Ryan Marten", "Mitchell Wortsman", "Dhruba Ghosh", "Jieyu Zhang", "Eyal Orgad", "Rahim Entezari", "Giannis Daras", "Sarah Pratt", "Vivek Ramanujan", "Yonatan Bitton", "Kalyani Marathe", "Stephen Mussmann", "Richard Vencu", "Mehdi Cherti", "Ranjay Krishna", "Pang Wei Koh", "Olga Saukh", "Alexander Ratner", "Shuran Song", "Hannaneh Hajishirzi", "Ali Farhadi", "Romain Beaumont", "Sewoong Oh", "Alexandros G. Dimakis", "Jenia Jitsev", "Yair Carmon", "Vaishaal Shankar", "Ludwig Schmidt" ], "authors_complete": true, "human_role_note": "Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.", "agents": [], "related": [], "subresults": [], "scientific_assessment": { "source_review": "SOURCE_CURATED", "scientific_review": "NOT_REVIEWED_BY_WRS", "formal_replay": "NOT_RUN", "provider_attestation": "NONE" }, "sources": [ { "id": "cur20261015-arxiv-2304-14108v1", "title": "DataComp: In search of the next generation of multimodal datasets", "url": "https://arxiv.org/abs/2304.14108v1" } ], "points": "115000", "published_grade": { "id": "gadre-et-al-datacomp-multimodal-benchmark-2023", "entry_id": "gadre-et-al-datacomp-multimodal-benchmark-2023", "band": "A", "modifier": null, "target": { "record_id": "urn:uuid:4ac30610-5039-4452-a733-8ee5804c5d78", "version": 1, "content_hash": "sha256:01b9bbd12090b79862f854137e563337499a2735850e77ce611f732dd734ca09" }, "what_changed": "Provided candidate image-text pools, fixed CLIP training/evaluation recipes and compute-scaled tracks, with DataComp-1B demonstrating the importance of filtering under controlled training conditions.", "scope": "Provided candidate image-text pools, fixed CLIP training/evaluation recipes and compute-scaled tracks, with DataComp-1B demonstrating the importance of filtering under controlled training conditions. Inherits CLIP and web-pair collection rather than inventing them. Ninefold compute comparison involves a larger LAION-trained architecture; not a universally hardware-matched data-only speedup. Fixed test suites do not guarantee uncontaminated or universal real-world generalization.", "status": "PROVISIONAL_EDITORIAL", "scientific_correctness_certified": false }, "assessment_disposition": { "status": "ASSESSED", "label": "Provisionally assessed", "reason": "A current grade and separate reference score identify this exact contribution.", "target": { "record_id": "urn:uuid:4ac30610-5039-4452-a733-8ee5804c5d78", "version": 1, "content_hash": "sha256:01b9bbd12090b79862f854137e563337499a2735850e77ce611f732dd734ca09" }, "edition": "ratings-e039106f30dec177" } }, "version": { "kind": "PUBLIC_RECORD_VIEW", "previous": null, "published_at": "2026-10-16T00:00:00Z", "publisher": { "principal_id": "urn:uuid:80116931-785b-40a2-a24b-9d59db0c7e33", "role": "WRS_EDITORIAL_PUBLICATION" }, "record": { "contributors": [ { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Samir Yitzhak Gadre", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DataComp: In search of the next generation of multimodal datasets", "url": "https://arxiv.org/abs/2304.14108v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Gabriel Ilharco", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DataComp: In search of the next generation of multimodal datasets", "url": "https://arxiv.org/abs/2304.14108v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Alex Fang", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DataComp: In search of the next generation of multimodal datasets", "url": "https://arxiv.org/abs/2304.14108v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Jonathan Hayase", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DataComp: In search of the next generation of multimodal datasets", "url": "https://arxiv.org/abs/2304.14108v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Georgios Smyrnis", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DataComp: In search of the next generation of multimodal datasets", "url": "https://arxiv.org/abs/2304.14108v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Thao Nguyen", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DataComp: In search of the next generation of multimodal datasets", "url": "https://arxiv.org/abs/2304.14108v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Ryan Marten", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DataComp: In search of the next generation of multimodal datasets", "url": "https://arxiv.org/abs/2304.14108v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Mitchell Wortsman", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DataComp: In search of the next generation of multimodal datasets", "url": "https://arxiv.org/abs/2304.14108v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Dhruba Ghosh", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DataComp: In search of the next generation of multimodal datasets", "url": "https://arxiv.org/abs/2304.14108v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Jieyu Zhang", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DataComp: In search of the next generation of multimodal datasets", "url": "https://arxiv.org/abs/2304.14108v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Eyal Orgad", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DataComp: In search of the next generation of multimodal datasets", "url": "https://arxiv.org/abs/2304.14108v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Rahim Entezari", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DataComp: In search of the next generation of multimodal datasets", "url": "https://arxiv.org/abs/2304.14108v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Giannis Daras", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DataComp: In search of the next generation of multimodal datasets", "url": "https://arxiv.org/abs/2304.14108v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Sarah Pratt", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DataComp: In search of the next generation of multimodal datasets", "url": "https://arxiv.org/abs/2304.14108v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Vivek Ramanujan", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DataComp: In search of the next generation of multimodal datasets", "url": "https://arxiv.org/abs/2304.14108v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Yonatan Bitton", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DataComp: In search of the next generation of multimodal datasets", "url": "https://arxiv.org/abs/2304.14108v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Kalyani Marathe", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DataComp: In search of the next generation of multimodal datasets", "url": "https://arxiv.org/abs/2304.14108v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Stephen Mussmann", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DataComp: In search of the next generation of multimodal datasets", "url": "https://arxiv.org/abs/2304.14108v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Richard Vencu", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DataComp: In search of the next generation of multimodal datasets", "url": "https://arxiv.org/abs/2304.14108v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Mehdi Cherti", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DataComp: In search of the next generation of multimodal datasets", "url": "https://arxiv.org/abs/2304.14108v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Ranjay Krishna", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DataComp: In search of the next generation of multimodal datasets", "url": "https://arxiv.org/abs/2304.14108v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Pang Wei Koh", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DataComp: In search of the next generation of multimodal datasets", "url": "https://arxiv.org/abs/2304.14108v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Olga Saukh", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DataComp: In search of the next generation of multimodal datasets", "url": "https://arxiv.org/abs/2304.14108v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Alexander Ratner", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DataComp: In search of the next generation of multimodal datasets", "url": "https://arxiv.org/abs/2304.14108v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Shuran Song", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DataComp: In search of the next generation of multimodal datasets", "url": "https://arxiv.org/abs/2304.14108v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Hannaneh Hajishirzi", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DataComp: In search of the next generation of multimodal datasets", "url": "https://arxiv.org/abs/2304.14108v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Ali Farhadi", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DataComp: In search of the next generation of multimodal datasets", "url": "https://arxiv.org/abs/2304.14108v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Romain Beaumont", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DataComp: In search of the next generation of multimodal datasets", "url": "https://arxiv.org/abs/2304.14108v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Sewoong Oh", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DataComp: In search of the next generation of multimodal datasets", "url": "https://arxiv.org/abs/2304.14108v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Alexandros G. Dimakis", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DataComp: In search of the next generation of multimodal datasets", "url": "https://arxiv.org/abs/2304.14108v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Jenia Jitsev", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "DataComp: In search of the next generation of multimodal datasets", "url": "https://arxiv.org/abs/2304.14108v1" } } ], "evidence": [ { "description": "Bibliographic/stated-result source; not a WRS proof verification", "id": "source-0", "kind": "source_reference", "source": { "citation": "DataComp: In search of the next generation of multimodal datasets", "url": "https://arxiv.org/abs/2304.14108v1" } } ], "origin": "HUMAN_LEGACY", "record_profile": "wrs-local-research-notes-0.1", "record_type": "research_note", "rrp_version": "0.1", "scope": { "description": "Inherits CLIP and web-pair collection rather than inventing them. Ninefold compute comparison involves a larger LAION-trained architecture; not a universally hardware-matched data-only speedup. Fixed test suites do not guarantee uncontaminated or universal real-world generalization.", "domain": "Multimodal dataset evaluation" }, "sources": [ { "citation": "DataComp: In search of the next generation of multimodal datasets", "url": "https://arxiv.org/abs/2304.14108v1" } ], "statement": "# DataComp: fixed-recipe multimodal dataset experiments\n\nProvided candidate image-text pools, fixed CLIP training/evaluation recipes and compute-scaled tracks, with DataComp-1B demonstrating the importance of filtering under controlled training conditions.\n\n## Scientific scope\n\nInherits CLIP and web-pair collection rather than inventing them. Ninefold compute comparison involves a larger LAION-trained architecture; not a universally hardware-matched data-only speedup. Fixed test suites do not guarantee uncontaminated or universal real-world generalization.\n\n## Reading status\n\nThis is a source-linked WRS catalogue note. A full explanatory guide and independent scientific reconstruction have not been completed for this entry. Read the identified original work for its complete narrative and arguments.\n\n## People and systems\n\nSamir Yitzhak Gadre; Gabriel Ilharco; Alex Fang; Jonathan Hayase; Georgios Smyrnis; Thao Nguyen; Ryan Marten; Mitchell Wortsman; Dhruba Ghosh; Jieyu Zhang; Eyal Orgad; Rahim Entezari; Giannis Daras; Sarah Pratt; Vivek Ramanujan; Yonatan Bitton; Kalyani Marathe; Stephen Mussmann; Richard Vencu; Mehdi Cherti; Ranjay Krishna; Pang Wei Koh; Olga Saukh; Alexander Ratner; Shuran Song; Hannaneh Hajishirzi; Ali Farhadi; Romain Beaumont; Sewoong Oh; Alexandros G. Dimakis; Jenia Jitsev; Yair Carmon; Vaishaal Shankar; Ludwig Schmidt\n\nSource-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.\n\n\n\n## Original sources\n\n- [DataComp: In search of the next generation of multimodal datasets](https://arxiv.org/abs/2304.14108v1)\n", "summary": "Provided candidate image-text pools, fixed CLIP training/evaluation recipes and compute-scaled tracks, with DataComp-1B demonstrating the importance of filtering under controlled training conditions.", "title": "DataComp: fixed-recipe multimodal dataset experiments" }, "record_id": "urn:uuid:4ac30610-5039-4452-a733-8ee5804c5d78", "version": 1, "source_ref": { "record_id": "urn:uuid:4ac30610-5039-4452-a733-8ee5804c5d78", "version": 1, "content_hash": "sha256:01b9bbd12090b79862f854137e563337499a2735850e77ce611f732dd734ca09" }, "export_view": "wrs-public-editorial-view-1", "public_view_hash": "sha256:bcd0ef773b3056ff5b7e0f97a74c39bbc9ece6d9568d40fc67ae51de465a4b87" }, "target": { "record_id": "urn:uuid:4ac30610-5039-4452-a733-8ee5804c5d78", "version": 1, "content_hash": "sha256:01b9bbd12090b79862f854137e563337499a2735850e77ce611f732dd734ca09" }, "published_grade": { "profile": "wrs-published-grade-1", "edition": "2026-10-28.1", "id": "gadre-et-al-datacomp-multimodal-benchmark-2023", "entry_id": "gadre-et-al-datacomp-multimodal-benchmark-2023", "revision": 1, "title": "DataComp: fixed-recipe multimodal dataset experiments", "context": "Multimodal dataset evaluation", "track": "Human-designed research method or resource", "band": "A", "modifier": null, "status": "PROVISIONAL_EDITORIAL", "assessed_on": "2026-10-16", "evidence_cutoff": "2026-10-16", "achievement_state": "SUPPORTED_SCOPED_RESULT", "what_changed": "Provided candidate image-text pools, fixed CLIP training/evaluation recipes and compute-scaled tracks, with DataComp-1B demonstrating the importance of filtering under controlled training conditions.", "scope": "Provided candidate image-text pools, fixed CLIP training/evaluation recipes and compute-scaled tracks, with DataComp-1B demonstrating the importance of filtering under controlled training conditions. Inherits CLIP and web-pair collection rather than inventing them. Ninefold compute comparison involves a larger LAION-trained architecture; not a universally hardware-matched data-only speedup. Fixed test suites do not guarantee uncontaminated or universal real-world generalization.", "evidence_summary": "The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Inherits CLIP and web-pair collection rather than inventing them. Ninefold compute comparison involves a larger LAION-trained architecture; not a universally hardware-matched data-only speedup. Fixed test suites do not guarantee uncontaminated or universal real-world generalization.", "references": [ { "title": "DataComp: In search of the next generation of multimodal datasets", "url": "https://arxiv.org/abs/2304.14108v1", "locator": "Original contribution and stated assumptions, evaluation scope and limitations; source-linked assessment, not independent execution or proof certification." } ], "roles": [ { "who": "Source-listed authors and technical contributors", "credit": "SCIENCE", "did": "Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities." } ], "target": { "record_id": "urn:uuid:4ac30610-5039-4452-a733-8ee5804c5d78", "version": 1, "content_hash": "sha256:01b9bbd12090b79862f854137e563337499a2735850e77ce611f732dd734ca09" }, "scientific_correctness_certified": false, "independent_expert_calibration": false, "method": "PROPRIETARY", "points": "115000" }, "canonical_agent_submission": false, "scientific_verification_awarded": false, "reader_guide": null, "reading_status": "SOURCE_LINKED_NOTE" }