{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:66ZBWGSAZNWNOG7K3IYNBIIHF7","short_pith_number":"pith:66ZBWGSA","schema_version":"1.0","canonical_sha256":"f7b21b1a40cb6cd71beada30d0a1072fdf468fea233e6f1494aa9420c4d828c1","source":{"kind":"arxiv","id":"2603.28767","version":3},"attestation_state":"computed","paper":{"title":"Gen-Searcher: Reinforcing Agentic Search for Image Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Gen-Searcher trains image generators to run multi-hop searches for external knowledge and reference images before synthesis.","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Chenyang Wang, Dian Zheng, Hongyu Li, Kaituo Feng, Kaixuan Fan, Manyuan Zhang, Shuang Chen, Xiangyu Yue, Yilei Jiang, Yunlong Lin","submitted_at":"2026-03-30T17:59:56Z","abstract_excerpt":"Recent image generation models have shown strong capabilities in generating high-fidelity and photorealistic images. However, they are fundamentally constrained by frozen internal knowledge, thus often failing on real-world scenarios that are knowledge-intensive or require up-to-date information. In this paper, we present Gen-Searcher, as the first attempt to train a search-augmented image generation agent, which performs multi-hop reasoning and search to collect the textual knowledge and reference images needed for grounded generation. To achieve this, we construct a tailored data pipeline an"},"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":"2603.28767","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-03-30T17:59:56Z","cross_cats_sorted":[],"title_canon_sha256":"57a41d644a8eb37996bb72f6f93a897936dc8ae53e3ae4cee6340580baacf3f5","abstract_canon_sha256":"9107c358b30541727fcc6adff8c7c3f755b0c01dfa522bc934431bc5fa819af5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-25T02:02:14.799402Z","signature_b64":"WU/h7fbcQjoVC8Sl0O0oC7HgM9Y/PJYBLmkKxVtr4MpLxpTm+8SvVEXIqjFP46IsfdAs8zk8J45T6NhHXtQ+Dw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f7b21b1a40cb6cd71beada30d0a1072fdf468fea233e6f1494aa9420c4d828c1","last_reissued_at":"2026-05-25T02:02:14.798740Z","signature_status":"signed_v1","first_computed_at":"2026-05-25T02:02:14.798740Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Gen-Searcher: Reinforcing Agentic Search for Image Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Gen-Searcher trains image generators to run multi-hop searches for external knowledge and reference images before synthesis.","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Chenyang Wang, Dian Zheng, Hongyu Li, Kaituo Feng, Kaixuan Fan, Manyuan Zhang, Shuang Chen, Xiangyu Yue, Yilei Jiang, Yunlong Lin","submitted_at":"2026-03-30T17:59:56Z","abstract_excerpt":"Recent image generation models have shown strong capabilities in generating high-fidelity and photorealistic images. However, they are fundamentally constrained by frozen internal knowledge, thus often failing on real-world scenarios that are knowledge-intensive or require up-to-date information. In this paper, we present Gen-Searcher, as the first attempt to train a search-augmented image generation agent, which performs multi-hop reasoning and search to collect the textual knowledge and reference images needed for grounded generation. To achieve this, we construct a tailored data pipeline an"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Experiments show that Gen-Searcher brings substantial gains, improving Qwen-Image by around 16 points on KnowGen and 15 points on WISE.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That multi-hop search reliably returns accurate, relevant textual knowledge and reference images that can be directly used for grounded generation without introducing new errors or noise.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Gen-Searcher is the first search-augmented image generation agent trained with SFT followed by agentic RL using dual text and image rewards on custom datasets and the KnowGen benchmark.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Gen-Searcher trains image generators to run multi-hop searches for external knowledge and reference images before synthesis.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"34783ca077761ecebae9fb5d17eb4681af2327799dc2dd9d9ef36732712179e5"},"source":{"id":"2603.28767","kind":"arxiv","version":3},"verdict":{"id":"7236a993-1409-4316-968b-65a87751c57b","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-14T21:14:52.265561Z","strongest_claim":"Experiments show that Gen-Searcher brings substantial gains, improving Qwen-Image by around 16 points on KnowGen and 15 points on WISE.","one_line_summary":"Gen-Searcher is the first search-augmented image generation agent trained with SFT followed by agentic RL using dual text and image rewards on custom datasets and the KnowGen benchmark.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That multi-hop search reliably returns accurate, relevant textual knowledge and reference images that can be directly used for grounded generation without introducing new errors or noise.","pith_extraction_headline":"Gen-Searcher trains image generators to run multi-hop searches for external knowledge and reference images before synthesis."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2603.28767/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":"2603.28767","created_at":"2026-05-25T02:02:14.798818+00:00"},{"alias_kind":"arxiv_version","alias_value":"2603.28767v3","created_at":"2026-05-25T02:02:14.798818+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2603.28767","created_at":"2026-05-25T02:02:14.798818+00:00"},{"alias_kind":"pith_short_12","alias_value":"66ZBWGSAZNWN","created_at":"2026-05-25T02:02:14.798818+00:00"},{"alias_kind":"pith_short_16","alias_value":"66ZBWGSAZNWNOG7K","created_at":"2026-05-25T02:02:14.798818+00:00"},{"alias_kind":"pith_short_8","alias_value":"66ZBWGSA","created_at":"2026-05-25T02:02:14.798818+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":9,"internal_anchor_count":9,"sample":[{"citing_arxiv_id":"2605.21605","citing_title":"GenEvolve: Self-Evolving Image Generation Agents via Tool-Orchestrated Visual Experience Distillation","ref_index":14,"is_internal_anchor":true},{"citing_arxiv_id":"2605.21605","citing_title":"GenEvolve: Self-Evolving Image Generation Agents via Tool-Orchestrated Visual Experience Distillation","ref_index":14,"is_internal_anchor":true},{"citing_arxiv_id":"2605.08063","citing_title":"Flow-OPD: On-Policy Distillation for Flow Matching Models","ref_index":13,"is_internal_anchor":true},{"citing_arxiv_id":"2605.18748","citing_title":"Aurora: Unified Video Editing with a Tool-Using Agent","ref_index":7,"is_internal_anchor":true},{"citing_arxiv_id":"2605.08063","citing_title":"Flow-OPD: On-Policy Distillation for Flow Matching Models","ref_index":13,"is_internal_anchor":true},{"citing_arxiv_id":"2605.12497","citing_title":"From Web to Pixels: Bringing Agentic Search into Visual Perception","ref_index":29,"is_internal_anchor":true},{"citing_arxiv_id":"2605.08063","citing_title":"Flow-OPD: On-Policy Distillation for Flow Matching Models","ref_index":13,"is_internal_anchor":true},{"citing_arxiv_id":"2605.08063","citing_title":"Flow-OPD: On-Policy Distillation for Flow Matching Models","ref_index":13,"is_internal_anchor":true},{"citing_arxiv_id":"2605.08043","citing_title":"SCOPE: Structured Decomposition and Conditional Skill Orchestration for Complex Image Generation","ref_index":28,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/66ZBWGSAZNWNOG7K3IYNBIIHF7","json":"https://pith.science/pith/66ZBWGSAZNWNOG7K3IYNBIIHF7.json","graph_json":"https://pith.science/api/pith-number/66ZBWGSAZNWNOG7K3IYNBIIHF7/graph.json","events_json":"https://pith.science/api/pith-number/66ZBWGSAZNWNOG7K3IYNBIIHF7/events.json","paper":"https://pith.science/paper/66ZBWGSA"},"agent_actions":{"view_html":"https://pith.science/pith/66ZBWGSAZNWNOG7K3IYNBIIHF7","download_json":"https://pith.science/pith/66ZBWGSAZNWNOG7K3IYNBIIHF7.json","view_paper":"https://pith.science/paper/66ZBWGSA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2603.28767&json=true","fetch_graph":"https://pith.science/api/pith-number/66ZBWGSAZNWNOG7K3IYNBIIHF7/graph.json","fetch_events":"https://pith.science/api/pith-number/66ZBWGSAZNWNOG7K3IYNBIIHF7/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/66ZBWGSAZNWNOG7K3IYNBIIHF7/action/timestamp_anchor","attest_storage":"https://pith.science/pith/66ZBWGSAZNWNOG7K3IYNBIIHF7/action/storage_attestation","attest_author":"https://pith.science/pith/66ZBWGSAZNWNOG7K3IYNBIIHF7/action/author_attestation","sign_citation":"https://pith.science/pith/66ZBWGSAZNWNOG7K3IYNBIIHF7/action/citation_signature","submit_replication":"https://pith.science/pith/66ZBWGSAZNWNOG7K3IYNBIIHF7/action/replication_record"}},"created_at":"2026-05-25T02:02:14.798818+00:00","updated_at":"2026-05-25T02:02:14.798818+00:00"}