{"ID":5936925,"CreatedAt":"2026-07-07T03:14:33.014478982Z","UpdatedAt":"2026-07-09T17:51:18.37832961Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2607.05382","arxiv_id":"2607.05382","title":"Search Beyond What Can Be Taught: Evolving the Knowledge Boundary in Agentic Visual Generation","abstract":"Visual generators excel at rendering, but they confidently fabricate what they do not know. User requests are unbounded, evolving, and deeply long-tailed: new characters, trending entities, post-cutoff events, and more. This world-knowledge bottleneck is structural: generators are trained on fixed corpora, but the visual world is open-ended. We construct SearchGen-20K and SearchGen-Bench, with 20,839 prompts spanning twelve failure categories and twenty-two domains, paired with a pre-executed multimodal SearchGen-Corpus-1M to support offline, reproducible research. On SearchGen-Bench, frontier open generators score only 21 to 28 out of 100, a 40-point collapse invisible to existing benchmarks. The natural remedy is to employ search tools, enabling agentic visual generation. However, we find that naive search fails: it retrieves indiscriminately, injecting noise into prompts the generator already handles. We trace the root cause to a generator-specific, evolving knowledge boundary: the divide between what a generator can internalize through training and what must remain in external context. Although this boundary is hard to specify in advance, we show that it is discoverable through a teach-then-search co-training framework. Even a minimal version of this co-training recipe produces monotonic improvement, laying the foundation for recursive self-improvement in visual generation that can meet world-knowledge-grounded requests. We release the full dataset, co-training corpus, and search corpus as a replayable harness for tool-augmented, world-knowledge-grounded visual generation.","short_abstract":"Visual generators excel at rendering, but they confidently fabricate what they do not know. User requests are unbounded, evolving, and deeply long-tailed: new characters, trending entities, post-cutoff events, and more. This world-knowledge bottleneck is structural: generators are trained on fixed corpora, but the visu...","url_abs":"https://arxiv.org/abs/2607.05382","url_pdf":"https://arxiv.org/pdf/2607.05382v1","authors":"[\"Haozhe Wang\",\"Weijia Feng\",\"Jinpeng Yu\",\"Che Liu\",\"Ping Nie\",\"Fangzhen Lin\",\"Jiaming Liu\",\"Ruihua Huang\",\"Jimmy Lin\",\"Wenhu Chen\",\"Cong Wei\"]","published":"2026-07-06T17:56:12Z","proceeding":"cs.CV","tasks":"[\"cs.CV\",\"cs.AI\"]","methods":"[]","has_code":false}