{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:C4Y7AZSJMKO4R3224MTDEDEZPK","short_pith_number":"pith:C4Y7AZSJ","schema_version":"1.0","canonical_sha256":"1731f06649629dc8ef5ae326320c997a8a32c786d3b5e010e293a7093eed99e6","source":{"kind":"arxiv","id":"2605.21059","version":1},"attestation_state":"computed","paper":{"title":"Multimodal LLMs under Pairwise Modalities","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Gongxu Luo, Guangyi Chen, Kun Zhang, Yan Li, Yuewen Sun, Yunlong Deng","submitted_at":"2026-05-20T11:44:01Z","abstract_excerpt":"Despite the impressive results achieved by multimodal large language models (MLLMs), their training typically relies on jointly curated multimodal data, requiring substantial human effort to construct multi-way aligned datasets and thereby limiting scalability across domains. In this work, we explore training MLLMs by only leveraging multiple paired modalities as a surrogate for the full joint multimodal distribution. Specifically, we first provide a theoretical analysis of the conditions under which the representations are identifiable with only observing pairwise modalities. Building on this"},"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":"2605.21059","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-20T11:44:01Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"4b95b7e687c367ad70596e69b5b3a6bbb00bcacfe91a0ef8706faef76df248ef","abstract_canon_sha256":"9316b96e5f03f1a74285a643f092b272949f20815692a9c927e9e3c79b677439"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-21T01:05:34.557578Z","signature_b64":"vc+XsQQM6B9aln/oc+mN2tL1TeMtK6+C0oWhJlmtj+s3u3KNnOBcGmip2+C5fOUc4OhRQ7HAO1JZIiSoVsuPBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1731f06649629dc8ef5ae326320c997a8a32c786d3b5e010e293a7093eed99e6","last_reissued_at":"2026-05-21T01:05:34.556371Z","signature_status":"signed_v1","first_computed_at":"2026-05-21T01:05:34.556371Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Multimodal LLMs under Pairwise Modalities","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Gongxu Luo, Guangyi Chen, Kun Zhang, Yan Li, Yuewen Sun, Yunlong Deng","submitted_at":"2026-05-20T11:44:01Z","abstract_excerpt":"Despite the impressive results achieved by multimodal large language models (MLLMs), their training typically relies on jointly curated multimodal data, requiring substantial human effort to construct multi-way aligned datasets and thereby limiting scalability across domains. In this work, we explore training MLLMs by only leveraging multiple paired modalities as a surrogate for the full joint multimodal distribution. Specifically, we first provide a theoretical analysis of the conditions under which the representations are identifiable with only observing pairwise modalities. Building on this"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.21059","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/2605.21059/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":"2605.21059","created_at":"2026-05-21T01:05:34.556537+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.21059v1","created_at":"2026-05-21T01:05:34.556537+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.21059","created_at":"2026-05-21T01:05:34.556537+00:00"},{"alias_kind":"pith_short_12","alias_value":"C4Y7AZSJMKO4","created_at":"2026-05-21T01:05:34.556537+00:00"},{"alias_kind":"pith_short_16","alias_value":"C4Y7AZSJMKO4R322","created_at":"2026-05-21T01:05:34.556537+00:00"},{"alias_kind":"pith_short_8","alias_value":"C4Y7AZSJ","created_at":"2026-05-21T01:05:34.556537+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/C4Y7AZSJMKO4R3224MTDEDEZPK","json":"https://pith.science/pith/C4Y7AZSJMKO4R3224MTDEDEZPK.json","graph_json":"https://pith.science/api/pith-number/C4Y7AZSJMKO4R3224MTDEDEZPK/graph.json","events_json":"https://pith.science/api/pith-number/C4Y7AZSJMKO4R3224MTDEDEZPK/events.json","paper":"https://pith.science/paper/C4Y7AZSJ"},"agent_actions":{"view_html":"https://pith.science/pith/C4Y7AZSJMKO4R3224MTDEDEZPK","download_json":"https://pith.science/pith/C4Y7AZSJMKO4R3224MTDEDEZPK.json","view_paper":"https://pith.science/paper/C4Y7AZSJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.21059&json=true","fetch_graph":"https://pith.science/api/pith-number/C4Y7AZSJMKO4R3224MTDEDEZPK/graph.json","fetch_events":"https://pith.science/api/pith-number/C4Y7AZSJMKO4R3224MTDEDEZPK/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/C4Y7AZSJMKO4R3224MTDEDEZPK/action/timestamp_anchor","attest_storage":"https://pith.science/pith/C4Y7AZSJMKO4R3224MTDEDEZPK/action/storage_attestation","attest_author":"https://pith.science/pith/C4Y7AZSJMKO4R3224MTDEDEZPK/action/author_attestation","sign_citation":"https://pith.science/pith/C4Y7AZSJMKO4R3224MTDEDEZPK/action/citation_signature","submit_replication":"https://pith.science/pith/C4Y7AZSJMKO4R3224MTDEDEZPK/action/replication_record"}},"created_at":"2026-05-21T01:05:34.556537+00:00","updated_at":"2026-05-21T01:05:34.556537+00:00"}