发表机构
Univ. Grenoble Alpes; CNRS; Inria; LIG; LJK（格勒诺布尔阿尔卑斯大学; 法国国家科学研究中心; 法国国家信息与自动化研究所; 格勒诺布尔信息学实验室; 格勒诺布尔数学与应用实验室）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本文提出利用熵正则最优传输（Sinkhorn散度）进行低预算主动学习中的核心集选择，在特征空间直接操作，获得无维度样本复杂度与高效梯度计算，实验优于现有启发式方法。
AI 中文摘要
我们考虑低预算主动学习，即选择有限数量的点（核心集），使得仅基于该选择就能训练出高精度的模型。在标签需要昂贵的专家干预的背景下（如医疗应用），这一问题尤为相关。我们利用从预训练的自监督模型中提取的特征来表示数据，并直接在特征空间中进行核心集选择。本文中，我们使用熵正则最优传输，特别是Sinkhorn散度，作为核心集选择标准，这首先使我们获得无维度依赖的样本复杂度结果，其次允许计算高效的梯度评估。这为使用基于梯度的算法快速计算候选解开辟了道路，并通过基于交换的局部搜索进一步改进，且对解的质量有保证。在图像基准和医疗数据集上的实验表明，我们的方法在低预算设置下优于最先进的启发式方法。
英文摘要
We consider low-budget active learning, which consists of selecting a limited number of points, the coreset, such that a model can be trained to high accuracy on the selection only. This problem is particularly relevant in contexts where labeling requires costly expert intervention, as in medical applications. We leverage features extracted from a pretrained self-supervised model to represent the data, and perform coreset selection directly in this feature space. In this paper, we use entropic optimal transport, specifically the Sinkhorn divergence, as the coreset selection criterion, which first allows us to get dimension-free sample complexity results, and second admits computationally efficient gradient evaluations. This opens the way to using gradient-based algorithms to rapidly compute solution candidates, further improved by a swap-based local search, with guarantees on the solution quality. Experiments on image benchmarks and medical datasets show that our method outperforms state-of-the-art heuristics in low-budget settings.
