Vendi Anomaly Score：从多样性视角做异常检测的新方法
Vendi Anomaly Scores for Efficient and Accurate Anomaly Detection
搞异常检测的同学可以看看这篇：VAS 不用调超参数，靠 Vendi 多样性分数的变化找异常，在 148 个基准数据集上跑了 SOTA。相比 Isolation Forest 这类老方法，理论更简洁，对大规模数据也更省心。
一篇 arXiv 论文提出 Vendi Anomaly Score（VAS），把异常检测重新表述为多样性问题：通过计算移除某个样本后数据集 Vendi 多样性分数的变化来判定异常。传统方法依赖局部密度、隔离启发式或分布拟合，对邻域尺度和分布假设敏感，VAS 则不需要密度估计和分区，也不需要超参数调参。
VAS 是非参数方法，计算量随数据集规模线性增长，能同时捕捉局部冗余和全局数据结构。作者在来自 10 个源数据集、4 种 embedding 架构的 148 个基准 embedding 数据集上测试，取得最先进的异常检测成绩。
论文还在大规模 ImageNet 实验中验证了方法的鲁棒性，随着污染率和数据规模增加，VAS 的表现依然稳定。对需要在科学数据或生产环境中做异常检测的工程师来说，这种免调参的特性降低了实际部署门槛。
Vendi Anomaly Scores for Efficient and Accurate Anomaly Detection
Anomaly detection (AD) is a critical for trustworthy machine learning and scientific discovery. Existing methods typically define anomalies through local density, isolation heuristics, or distance from a fitted distribution, making them sensitive to neighborhood scale, partitioning choices, or restrictive distributional assumptions. In this work, we introduce a new paradigm by formulating anomaly detection from a diversity perspective. We propose the $\textit{Vendi Anomaly Score}$ (VAS), which detects anomalies by quantifying how much a sample changes the diversity of the dataset, as measured by the Vendi score, when removed. VAS captures both local redundancy and global data structure without relying on density estimation or partitioning heuristics. Moreover, VAS requires no hyperparameter tuning, instead adapting automatically to the spectral structure of the data. VAS is non-parametric and scales linearly with dataset size. Across 148 benchmark embedding datasets derived from 10 source datasets and 4 embedding architectures, VAS achieves state-of-the-art AD performance. We further validate VAS on large-scale ImageNet experiments, where it remains robust as both contamination rate and dataset scale increase.
本条情报归属 AITOP「论文」核心追踪栏目。涉及核心实体与技术议题包括：异常检测、Vendi score、ImageNet、机器学习，读者可通过下方标签进入对应专题页查阅历史脉络与关联对比。
信源核验说明：本文原始线索采编自「arXiv cs.LG」（@Amey Pasarkar, Adji Bousso Dieng），经 AITOP 多维评分管线结构化提炼。