发表机构
Stevens Institute of Technology（史蒂文斯理工学院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本文提出ReMaD方法，利用降秩马氏距离对预训练模型进行无需微调的域自适应，在分类和分布外检测中取得竞争性能。
AI 中文摘要
我们引入了降秩马氏距离（ReMaD），一种新颖的基于原型距离的细化方法，用于使用预训练模型进行分类和分布外（OOD）检测，无需微调。我们利用目标数据集的嵌入来拟合模型潜在空间中的闭式分布统计量，从而能够对分布内样本进行分类并检测OOD样本，所有这些都不需要训练或对OOD数据的先验知识。基于原型分类和OOD检测，我们分析了大型预训练模型在处理新数据集时的分布特性；基于此分析，我们提出了对马氏距离的简单修改，通过移除未使用的特征来调整模型潜在空间分布以适应新域，无需其他自适应过程所需的微调或超参数搜索。我们通过在四个目标数据集上进行测试，证明了该方法能够将现有的大型预训练图像嵌入模型适应其训练能力之外的新分类域，并在分类和OOD检测中均展现出具有竞争力的性能。
英文摘要
We introduce Reduced-rank Mahalanobis Distance (ReMaD), a novel prototypical distance-based refinement to classification and out-of-distribution (OOD) detection using pretrained models without finetuning. We use embeddings of the target dataset to fit closed-form distribution statistics in the model's latent space which can classify in-distribution samples and detect OOD samples, all without training or prior knowledge of the OOD data. Building on prototype classification and OOD detection, we analyze the distribution properties of large pretrained models when processing new datasets; based on this analysis, we formulate a simple modification to Mahalanobis Distance to adapt models' latent space distributions to new domains by removing unused features, without the finetuning or hyperparameter searches required by other adaptation procedures. We demonstrate the efficacy of this method to adapt existing large pretrained image embedding models to new classification domains outside their trained capabilities by testing across four target datasets, with competitive performance in both classification and OOD detection.
