Search Beyond What Can Be Taught: Evolving the Knowledge Boundary in Agentic Visual Generation
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
Lineage graph
Paper → model → repo connections mined from source citations (Tier-1 exact match).
Why these links exist
Every edge carries a method, confidence, and the source snippet that justified it — so bad links are debuggable.
- PossiblePossibly related (embedding) · 56%llmsresearch/llm-flashcards →
- PossiblePossibly related (embedding) · 51%xlang-ai/OSWorld →
- PossiblePossibly related (embedding) · 51%stackitcloud/rag-template →
- PossiblePossibly related (embedding) · 51%apache/texera →
- PossiblePossibly related (embedding) · 51%ghbalf/freecad-ai →
- LinkedLinked via arxiv author · 85%Haozhe Wang →
“Search Beyond What Can Be Taught: Evolving the Knowledge Boundary in Agentic Visual Generation”
- LinkedLinked via arxiv author · 85%Weijia Feng →
“Search Beyond What Can Be Taught: Evolving the Knowledge Boundary in Agentic Visual Generation”
- LinkedLinked via arxiv author · 85%Jinpeng Yu →
“Search Beyond What Can Be Taught: Evolving the Knowledge Boundary in Agentic Visual Generation”