SPoRC-VIST：评估视觉语言模型生成自然叙述能力的基准
SPoRC-VIST: A Benchmark for Evaluating Generative Natural Narrative in Vision-Language Models
AI总结：
SPoRC-VIST基准通过合成到现实训练策略评估视觉语言模型生成多说话者播客对话的自然性和深度。
摘要：
视觉语言模型（VLMs）在图像描述任务如图像描述和视觉问答（VQA）中取得了显著成功。然而，它们生成吸引人、长篇叙述的能力——特别是多说话者播客对话——仍然未被充分探索和评估。标准度量如BLEU和ROUGE无法捕捉对话自然性、个性和叙述流畅性的细微差别，往往奖励安全、重复的输出而非引人入胜的故事叙述。在本工作中，我们提出了一种端到端的视觉播客生成流程，并在精心编纂的4000对图像-对话数据集上微调了Qwen3-VL-32B模型。关键在于我们采用了合成到现实的训练策略：我们使用高质量的播客对话从结构化播客研究语料库（SPoRC）配对合成图像进行训练，并在现实世界的视觉叙事数据集（VIST）上进行评估。这种严格的方法测试了模型从合成训练数据到真实视觉领域的泛化能力。我们提出了一种全面的评估框架，超越了文本重叠，并使用AI作为评判者（Gemini 3 Pro、Claude Opus 4.5、GPT 5.2）和新的风格度量（平均轮次长度、说话人切换率）来评估质量。我们的实验表明，微调后的32B模型在对话自然性（胜率超过80%）和叙述深度（轮次长度增加50%）方面显著优于235B基础模型，同时保持了相同的视觉定位能力（CLIPScore：20.39）。
英文摘要：
Vision-Language Models (VLMs) have achieved remarkable success in descriptive tasks such as image captioning and visual question answering (VQA). However, their ability to generate engaging, long-form narratives -- specifically multi-speaker podcast dialogues -- remains under-explored and difficult to evaluate. Standard metrics like BLEU and ROUGE fail to capture the nuances of conversational naturalness, personality, and narrative flow, often rewarding safe, repetitive outputs over engaging storytelling. In this work, we present a novel pipeline for end-to-end visual podcast generation, and fine-tune a Qwen3-VL-32B model on a curated dataset of 4,000 image-dialogue pairs. Crucially, we use a synthetic-to-real training strategy: we train on high-quality podcast dialogues from the Structured Podcast Research Corpus (SPoRC) paired with synthetically generated imagery, and evaluate on real-world photo sequences from the Visual Storytelling Dataset (VIST). This rigorous setup tests the model's ability to generalize from synthetic training data to real-world visual domains. We propose a comprehensive evaluation framework that moves beyond textual overlap, and use AI-as-a-judge (Gemini 3 Pro, Claude Opus 4.5, GPT 5.2) and novel style metrics (average turn length, speaker switch rate) to assess quality. Our experiments demonstrate that our fine-tuned 32B model significantly outperforms a 235B base model in conversational naturalness ($>$80\% win rate) and narrative depth (+50\% turn length), while maintaining identical visual grounding capabilities (CLIPScore: 20.39).
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
