{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:GUK4JKAGNISNGJ2GE5IPHO5BWU","short_pith_number":"pith:GUK4JKAG","schema_version":"1.0","canonical_sha256":"3515c4a8066a24d327462750f3bba1b51cc07756051c927c85d4954bfb0f12f9","source":{"kind":"arxiv","id":"2503.19312","version":1},"attestation_state":"computed","paper":{"title":"ImageGen-CoT: Enhancing Text-to-Image In-context Learning with Chain-of-Thought Reasoning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Dianqi Li, Jiaqi Liao, Kevin Lin, Lijuan Wang, Linjie Li, Yu Cheng, Zhengyuan Yang","submitted_at":"2025-03-25T03:18:46Z","abstract_excerpt":"In this work, we study the problem of Text-to-Image In-Context Learning (T2I-ICL). While Unified Multimodal LLMs (MLLMs) have advanced rapidly in recent years, they struggle with contextual reasoning in T2I-ICL scenarios. To address this limitation, we propose a novel framework that incorporates a thought process called ImageGen-CoT prior to image generation. To avoid generating unstructured ineffective reasoning steps, we develop an automatic pipeline to curate a high-quality ImageGen-CoT dataset. We then fine-tune MLLMs using this dataset to enhance their contextual reasoning capabilities. T"},"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":"2503.19312","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-03-25T03:18:46Z","cross_cats_sorted":[],"title_canon_sha256":"19950665b6f24c201066dc5a358f3e17ce3ea35e0856207850d95032b18424fb","abstract_canon_sha256":"44765dec9978ca7780e809cb38e66480f2c1be9788d438a672aa8c345919117d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:38:53.043519Z","signature_b64":"CmZEztc+7HhDd55EkrFavXcGPLG29vcCkaOSNuJosuo22HK4YL5cz/Q+2IQX3ax+c+W/kNgfGRQWtC66gFq0Cg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3515c4a8066a24d327462750f3bba1b51cc07756051c927c85d4954bfb0f12f9","last_reissued_at":"2026-07-05T10:38:53.043068Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:38:53.043068Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"ImageGen-CoT: Enhancing Text-to-Image In-context Learning with Chain-of-Thought Reasoning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Dianqi Li, Jiaqi Liao, Kevin Lin, Lijuan Wang, Linjie Li, Yu Cheng, Zhengyuan Yang","submitted_at":"2025-03-25T03:18:46Z","abstract_excerpt":"In this work, we study the problem of Text-to-Image In-Context Learning (T2I-ICL). While Unified Multimodal LLMs (MLLMs) have advanced rapidly in recent years, they struggle with contextual reasoning in T2I-ICL scenarios. To address this limitation, we propose a novel framework that incorporates a thought process called ImageGen-CoT prior to image generation. To avoid generating unstructured ineffective reasoning steps, we develop an automatic pipeline to curate a high-quality ImageGen-CoT dataset. We then fine-tune MLLMs using this dataset to enhance their contextual reasoning capabilities. T"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.19312","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/2503.19312/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":"2503.19312","created_at":"2026-07-05T10:38:53.043126+00:00"},{"alias_kind":"arxiv_version","alias_value":"2503.19312v1","created_at":"2026-07-05T10:38:53.043126+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.19312","created_at":"2026-07-05T10:38:53.043126+00:00"},{"alias_kind":"pith_short_12","alias_value":"GUK4JKAGNISN","created_at":"2026-07-05T10:38:53.043126+00:00"},{"alias_kind":"pith_short_16","alias_value":"GUK4JKAGNISNGJ2G","created_at":"2026-07-05T10:38:53.043126+00:00"},{"alias_kind":"pith_short_8","alias_value":"GUK4JKAG","created_at":"2026-07-05T10:38:53.043126+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.30124","citing_title":"SciIR: A Large-scale Training Dataset and Benchmark for Scientific Image Reasoning Generation","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2605.14876","citing_title":"Unlocking Complex Visual Generation via Closed-Loop Verified Reasoning","ref_index":26,"is_internal_anchor":false},{"citing_arxiv_id":"2512.22437","citing_title":"EmoCtrl: Controllable Emotional Image Content Generation","ref_index":17,"is_internal_anchor":false},{"citing_arxiv_id":"2604.02355","citing_title":"From Broad Exploration to Stable Synthesis: Entropy-Guided Optimization for Autoregressive Image Generation","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2604.25477","citing_title":"DDA-Thinker: Decoupled Dual-Atomic Reinforcement Learning for Reasoning-Driven Image Editing","ref_index":12,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/GUK4JKAGNISNGJ2GE5IPHO5BWU","json":"https://pith.science/pith/GUK4JKAGNISNGJ2GE5IPHO5BWU.json","graph_json":"https://pith.science/api/pith-number/GUK4JKAGNISNGJ2GE5IPHO5BWU/graph.json","events_json":"https://pith.science/api/pith-number/GUK4JKAGNISNGJ2GE5IPHO5BWU/events.json","paper":"https://pith.science/paper/GUK4JKAG"},"agent_actions":{"view_html":"https://pith.science/pith/GUK4JKAGNISNGJ2GE5IPHO5BWU","download_json":"https://pith.science/pith/GUK4JKAGNISNGJ2GE5IPHO5BWU.json","view_paper":"https://pith.science/paper/GUK4JKAG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2503.19312&json=true","fetch_graph":"https://pith.science/api/pith-number/GUK4JKAGNISNGJ2GE5IPHO5BWU/graph.json","fetch_events":"https://pith.science/api/pith-number/GUK4JKAGNISNGJ2GE5IPHO5BWU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/GUK4JKAGNISNGJ2GE5IPHO5BWU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/GUK4JKAGNISNGJ2GE5IPHO5BWU/action/storage_attestation","attest_author":"https://pith.science/pith/GUK4JKAGNISNGJ2GE5IPHO5BWU/action/author_attestation","sign_citation":"https://pith.science/pith/GUK4JKAGNISNGJ2GE5IPHO5BWU/action/citation_signature","submit_replication":"https://pith.science/pith/GUK4JKAGNISNGJ2GE5IPHO5BWU/action/replication_record"}},"created_at":"2026-07-05T10:38:53.043126+00:00","updated_at":"2026-07-05T10:38:53.043126+00:00"}