提出条件化多样性评估方法，精准区分提示与模型带来的生成差异。
Conditional Vendi Score: Prompt-Aware Diversity Evaluation for Generative AI Models and LLMs
- 基于条件熵构建新度量，分离提示与模型引发的多样性
- 在文本到图像等任务中恢复真实多样性排序，提升扩散模型采样多样性
- 适用于需控制生成多样性的生成模型与大模型评估
由文本提示引导的生成模型虽广泛评估保真度与提示对齐性，但其生成能力中的多样性仍缺乏深入研究。现有基于冯诺依曼与黎曼熵的无条件多样性度量（如 Vendi、RKE）无法区分提示引发与模型自身导致的变异性。本文提出 extit{Conditional-Vendi} 与 extit{Conditional-RKE}，从半正定矩阵的条件熵出发，隔离出提示引导下模型自身的多样性。其中，Conditional-RKE 具有 $O(1/ ext{sqrt}{n})$ 的收敛速率；针对 Conditional-Vendi，引入截断谱近似以实现可扩展且一致的估计。在文本到图像、图像描述及大模型任务上的实验表明，该方法能准确恢复真实多样性排序，并有效指导扩散模型生成更丰富样本。代码已开源：https://github.com/mjalali/conditional-vendi。
原文摘要 · Abstract (English)
Generative models guided by text prompts are widely evaluated for fidelity and prompt alignment, yet their ability to produce outputs remains underexplored. Existing diversity metrics such as Vendi and RKE, which are based on the von Neumann and Rényi entropies of kernel matrices, were developed for unconditional models and cannot distinguish prompt-induced from model-induced variability. We address this gap by introducing \textit{Conditional-Vendi} and \textit{Conditional-RKE}, diversity measures derived from the conditional entropy of positive semidefinite matrices. These scores isolate model-induced diversity in prompt-guided generation, with Conditional-RKE enjoying an $O(1/\sqrt{n})$ convergence rate. For Conditional-Vendi, we introduce a truncated-spectrum approximation that yields scalable and consistent estimates. Experiments on text-to-image, image-captioning, and LLM tasks show that the conditional scores recover ground-truth diversity orderings and can also guide diffusion models toward more diverse samples. The codebase is available at https://github.com/mjalali/conditional-vendi.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。