VQ-VA World：迈向高质量视觉问题视觉回答
VQ-VA World: Towards High-Quality Visual Question-Visual Answering
论文提出以图像而非文本回答视觉问题的 VQ-VA 任务，并给出基于智能体流水线的大规模交错图文数据构建方案与 IntelligentBench 基准，为视觉语言模型从判别式问答向生成式视觉回答扩展提供了可复用的数据与评测范式。
原文依据 This paper studies Visual Question-Visual Answering (VQ-VA): generating an image, rather than text, in response to a visual question
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论本文研究视觉问题视觉回答（VQ-VA）：针对视觉问题生成图像而非文本——这一能力近期已在 NanoBanana 和 GPT-Image 等专有系统中出现。为将这一能力也带给开源模型，我们提出 VQ-VA World，一个以数据为中心、围绕智能体流水线构建的框架，用于大规模、有针对性的数据构建。借助网络级部署，该流水线爬取了约 180 万条高质量交错图文样本用于模型训练。在评估方面，我们进一步发布 IntelligentBench，一个经人工整理的基准，从世界知识、设计知识和推理三个方面系统评估 VQ-VA。使用 VQ-VA World 数据训练带来显著的实证增益：它帮助 LightFusion 在 IntelligentBench 上达到 53.06，大幅超越最佳开源基线（即原始 LightFusion 的 7.78；UniWorld-V1 的 1.94），并显著缩小与领先专有系统的差距（例如 NanoBanana 的 81.67；GPT-Image 的 82.64）。通过发布全套模型权重、数据集和流水线，我们的工作将极大促进未来 VQ-VA 研究。项目已公开于 https://github.com/chenhuigou/VQ-VA-World。
英文原文摘要 1314 字符
This paper studies Visual Question-Visual Answering (VQ-VA): generating an image, rather than text, in response to a visual question---an ability that has recently emerged in proprietary systems such as NanoBanana and GPT-Image. To also bring this capability to open-source models, we introduce VQ-VA World, a data-centric framework built around an agentic pipeline for large-scale, targeted data construction. Leveraging web-scale deployment, this pipeline crawls a massive amount of ~1.8M high-quality, interleaved image-text samples for model training. For evaluation, we further release IntelligentBench, a human-curated benchmark that systematically assesses VQ-VA along the aspects of world knowledge, design knowledge, and reasoning. Training with VQ-VA World data yields strong empirical gains: it helps LightFusion attain 53.06 on IntelligentBench, substantially surpassing the best prior open-source baselines (i.e., 7.78 from vanilla LightFusion; 1.94 from UniWorld-V1), and significantly narrowing the gap toward leading proprietary systems (e.g., 81.67 from NanoBanana; 82.64 from GPT-Image). By releasing the full suite of model weights, datasets, and pipelines, our work will greatly stimulate future research on VQ-VA. Our project is publicly available at https://github.com/chenhuigou/VQ-VA-World.