{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:CYYK37H7UWKHUPU7MG4J5YABTX","short_pith_number":"pith:CYYK37H7","schema_version":"1.0","canonical_sha256":"1630adfcffa5947a3e9f61b89ee0019dc62b40a46fe2dc39fa31885cfc05dad2","source":{"kind":"arxiv","id":"2502.10458","version":1},"attestation_state":"computed","paper":{"title":"I Think, Therefore I Diffuse: Enabling Multimodal In-Context Reasoning in Diffusion Models","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Dan Xu, Guocheng Qian, Hanrong Ye, Kfir Aberman, Kuan-Chieh Wang, Runtao Liu, Sergey Tulyakov, Zhenxing Mi","submitted_at":"2025-02-12T05:30:08Z","abstract_excerpt":"This paper presents ThinkDiff, a novel alignment paradigm that empowers text-to-image diffusion models with multimodal in-context understanding and reasoning capabilities by integrating the strengths of vision-language models (VLMs). Existing multimodal diffusion finetuning methods largely focus on pixel-level reconstruction rather than in-context reasoning, and are constrained by the complexity and limited availability of reasoning-based datasets. ThinkDiff addresses these challenges by leveraging vision-language training as a proxy task, aligning VLMs with the decoder of an encoder-decoder l"},"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":"2502.10458","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-12T05:30:08Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"23087d0f96ced5c1a02d99092138fa080d63e873fb5b0dc1e9a261de6c0def9b","abstract_canon_sha256":"d6fe043b426db740d6fcf30b5b9febcebb1a3cd41b31f668b8d181007c330cc9"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:14:52.972499Z","signature_b64":"DDxY+L65o3v/Mqc7HkcYpBJaYIHZgbIOH7MgAurvUQZh4jbpdoFbXRZHPB7F75Np4MYgsDVN1+DypLjr/RIiDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1630adfcffa5947a3e9f61b89ee0019dc62b40a46fe2dc39fa31885cfc05dad2","last_reissued_at":"2026-07-05T10:14:52.972020Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:14:52.972020Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"I Think, Therefore I Diffuse: Enabling Multimodal In-Context Reasoning in Diffusion Models","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Dan Xu, Guocheng Qian, Hanrong Ye, Kfir Aberman, Kuan-Chieh Wang, Runtao Liu, Sergey Tulyakov, Zhenxing Mi","submitted_at":"2025-02-12T05:30:08Z","abstract_excerpt":"This paper presents ThinkDiff, a novel alignment paradigm that empowers text-to-image diffusion models with multimodal in-context understanding and reasoning capabilities by integrating the strengths of vision-language models (VLMs). Existing multimodal diffusion finetuning methods largely focus on pixel-level reconstruction rather than in-context reasoning, and are constrained by the complexity and limited availability of reasoning-based datasets. ThinkDiff addresses these challenges by leveraging vision-language training as a proxy task, aligning VLMs with the decoder of an encoder-decoder l"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.10458","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/2502.10458/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":"2502.10458","created_at":"2026-07-05T10:14:52.972077+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.10458v1","created_at":"2026-07-05T10:14:52.972077+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.10458","created_at":"2026-07-05T10:14:52.972077+00:00"},{"alias_kind":"pith_short_12","alias_value":"CYYK37H7UWKH","created_at":"2026-07-05T10:14:52.972077+00:00"},{"alias_kind":"pith_short_16","alias_value":"CYYK37H7UWKHUPU7","created_at":"2026-07-05T10:14:52.972077+00:00"},{"alias_kind":"pith_short_8","alias_value":"CYYK37H7","created_at":"2026-07-05T10:14:52.972077+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.04479","citing_title":"Evaluating Reasoning Fidelity in Visual Text Generation","ref_index":32,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/CYYK37H7UWKHUPU7MG4J5YABTX","json":"https://pith.science/pith/CYYK37H7UWKHUPU7MG4J5YABTX.json","graph_json":"https://pith.science/api/pith-number/CYYK37H7UWKHUPU7MG4J5YABTX/graph.json","events_json":"https://pith.science/api/pith-number/CYYK37H7UWKHUPU7MG4J5YABTX/events.json","paper":"https://pith.science/paper/CYYK37H7"},"agent_actions":{"view_html":"https://pith.science/pith/CYYK37H7UWKHUPU7MG4J5YABTX","download_json":"https://pith.science/pith/CYYK37H7UWKHUPU7MG4J5YABTX.json","view_paper":"https://pith.science/paper/CYYK37H7","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.10458&json=true","fetch_graph":"https://pith.science/api/pith-number/CYYK37H7UWKHUPU7MG4J5YABTX/graph.json","fetch_events":"https://pith.science/api/pith-number/CYYK37H7UWKHUPU7MG4J5YABTX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CYYK37H7UWKHUPU7MG4J5YABTX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CYYK37H7UWKHUPU7MG4J5YABTX/action/storage_attestation","attest_author":"https://pith.science/pith/CYYK37H7UWKHUPU7MG4J5YABTX/action/author_attestation","sign_citation":"https://pith.science/pith/CYYK37H7UWKHUPU7MG4J5YABTX/action/citation_signature","submit_replication":"https://pith.science/pith/CYYK37H7UWKHUPU7MG4J5YABTX/action/replication_record"}},"created_at":"2026-07-05T10:14:52.972077+00:00","updated_at":"2026-07-05T10:14:52.972077+00:00"}