论文标题
SOQAL：主动学习中的选择性甲骨文提问
SoQal: Selective Oracle Questioning in Active Learning
论文作者
论文摘要
医疗领域内的大量未标记数据仍然不足。主动学习提供了一种通过迭代要求Oracle（例如医学专家）标记实例来利用这些数据集的方法。这个过程可能是昂贵的，耗时的过程过于依赖于甲骨文。为了减轻这一负担，我们提出了Soqal，这是一种质疑策略，该策略会动态地确定何时应从Oracle请求标签。我们对五个公开可用的数据集进行了实验，并说明了Soqal相对于基线方法的优势，包括其将Oracle标签请求最多减少35％的能力。 Soqal在标签噪声的存在下还具有竞争力：在面临困难的分类任务时模拟临床医生不确定的诊断的情况。
Large sets of unlabelled data within the healthcare domain remain underutilized. Active learning offers a way to exploit these datasets by iteratively requesting an oracle (e.g. medical professional) to label instances. This process, which can be costly and time-consuming is overly-dependent upon an oracle. To alleviate this burden, we propose SoQal, a questioning strategy that dynamically determines when a label should be requested from an oracle. We perform experiments on five publically-available datasets and illustrate SoQal's superiority relative to baseline approaches, including its ability to reduce oracle label requests by up to 35%. SoQal also performs competitively in the presence of label noise: a scenario that simulates clinicians' uncertain diagnoses when faced with difficult classification tasks.
