{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:M34V7ZSBXJBDJVYVELOX6RWQKQ","short_pith_number":"pith:M34V7ZSB","schema_version":"1.0","canonical_sha256":"66f95fe641ba4234d71522dd7f46d0540603c06d567f46461a5e2cc502e5ee04","source":{"kind":"arxiv","id":"2401.00849","version":1},"attestation_state":"computed","paper":{"title":"COSMO: COntrastive Streamlined MultimOdal Model with Interleaved Pre-Training","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Alex Jinpeng Wang, Jianfeng Wang, Kevin Lin, Kevin Qinghong Lin, Lijuan Wang, Linjie Li, Mike Zheng Shou, Zhengyuan Yang","submitted_at":"2024-01-01T18:58:42Z","abstract_excerpt":"In the evolution of Vision-Language Pre-training, shifting from short-text comprehension to encompassing extended textual contexts is pivotal. Recent autoregressive vision-language models like \\cite{flamingo, palme}, leveraging the long-context capability of Large Language Models, have excelled in few-shot text generation tasks but face challenges in alignment tasks. Addressing this gap, we introduce the contrastive loss into text generation models, presenting the COntrastive-Streamlined MultimOdal framework (\\ModelName), strategically partitioning the language model into dedicated unimodal te"},"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":"2401.00849","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2024-01-01T18:58:42Z","cross_cats_sorted":[],"title_canon_sha256":"dfbb8c89626c50d94645f2b459936458baa7b6577b96fd8b9e69461593b1e681","abstract_canon_sha256":"6942303525929f855833d75ff8d6c09f7c961600c0bc19dd0fd32c2d6756e247"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:29:23.596753Z","signature_b64":"7cEFYEIwfYAyPYS3AOq9ICNJZt1EPEsngel+y0k2X/rs8chOTLA7//Eofpymn9c3m5Ih6sQt+4+EiScf7uPmDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"66f95fe641ba4234d71522dd7f46d0540603c06d567f46461a5e2cc502e5ee04","last_reissued_at":"2026-07-05T07:29:23.596211Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:29:23.596211Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"COSMO: COntrastive Streamlined MultimOdal Model with Interleaved Pre-Training","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Alex Jinpeng Wang, Jianfeng Wang, Kevin Lin, Kevin Qinghong Lin, Lijuan Wang, Linjie Li, Mike Zheng Shou, Zhengyuan Yang","submitted_at":"2024-01-01T18:58:42Z","abstract_excerpt":"In the evolution of Vision-Language Pre-training, shifting from short-text comprehension to encompassing extended textual contexts is pivotal. Recent autoregressive vision-language models like \\cite{flamingo, palme}, leveraging the long-context capability of Large Language Models, have excelled in few-shot text generation tasks but face challenges in alignment tasks. Addressing this gap, we introduce the contrastive loss into text generation models, presenting the COntrastive-Streamlined MultimOdal framework (\\ModelName), strategically partitioning the language model into dedicated unimodal te"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2401.00849","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/2401.00849/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":"2401.00849","created_at":"2026-07-05T07:29:23.596281+00:00"},{"alias_kind":"arxiv_version","alias_value":"2401.00849v1","created_at":"2026-07-05T07:29:23.596281+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2401.00849","created_at":"2026-07-05T07:29:23.596281+00:00"},{"alias_kind":"pith_short_12","alias_value":"M34V7ZSBXJBD","created_at":"2026-07-05T07:29:23.596281+00:00"},{"alias_kind":"pith_short_16","alias_value":"M34V7ZSBXJBDJVYV","created_at":"2026-07-05T07:29:23.596281+00:00"},{"alias_kind":"pith_short_8","alias_value":"M34V7ZSB","created_at":"2026-07-05T07:29:23.596281+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":6,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.12195","citing_title":"InternVideo3: Agentify Foundation Models with Multimodal Contextual Reasoning","ref_index":281,"is_internal_anchor":false},{"citing_arxiv_id":"2606.06294","citing_title":"Towards One-to-Many Temporal Grounding","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2606.05758","citing_title":"DRIFT: A Residual Flow Adapter for Decoding Continuous Outputs in Vision-Language Models","ref_index":54,"is_internal_anchor":false},{"citing_arxiv_id":"2501.00574","citing_title":"VideoChat-Flash: Hierarchical Compression for Long-Context Video Modeling","ref_index":51,"is_internal_anchor":false},{"citing_arxiv_id":"2511.04670","citing_title":"Cambrian-S: Towards Spatial Supersensing in Video","ref_index":132,"is_internal_anchor":false},{"citing_arxiv_id":"2604.11283","citing_title":"Multimodal Large Language Model-Enabled Video Translation: A Role-Oriented Survey","ref_index":13,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/M34V7ZSBXJBDJVYVELOX6RWQKQ","json":"https://pith.science/pith/M34V7ZSBXJBDJVYVELOX6RWQKQ.json","graph_json":"https://pith.science/api/pith-number/M34V7ZSBXJBDJVYVELOX6RWQKQ/graph.json","events_json":"https://pith.science/api/pith-number/M34V7ZSBXJBDJVYVELOX6RWQKQ/events.json","paper":"https://pith.science/paper/M34V7ZSB"},"agent_actions":{"view_html":"https://pith.science/pith/M34V7ZSBXJBDJVYVELOX6RWQKQ","download_json":"https://pith.science/pith/M34V7ZSBXJBDJVYVELOX6RWQKQ.json","view_paper":"https://pith.science/paper/M34V7ZSB","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2401.00849&json=true","fetch_graph":"https://pith.science/api/pith-number/M34V7ZSBXJBDJVYVELOX6RWQKQ/graph.json","fetch_events":"https://pith.science/api/pith-number/M34V7ZSBXJBDJVYVELOX6RWQKQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/M34V7ZSBXJBDJVYVELOX6RWQKQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/M34V7ZSBXJBDJVYVELOX6RWQKQ/action/storage_attestation","attest_author":"https://pith.science/pith/M34V7ZSBXJBDJVYVELOX6RWQKQ/action/author_attestation","sign_citation":"https://pith.science/pith/M34V7ZSBXJBDJVYVELOX6RWQKQ/action/citation_signature","submit_replication":"https://pith.science/pith/M34V7ZSBXJBDJVYVELOX6RWQKQ/action/replication_record"}},"created_at":"2026-07-05T07:29:23.596281+00:00","updated_at":"2026-07-05T07:29:23.596281+00:00"}