arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2008.10548cs.CV
收藏

多实例学习中的安全池

Certainty Pooling for Multiple Instance Learning

SagivTech Ltd.(SagivTech有限公司);Roche Innovation Center Munich(罗氏创新中心慕尼黑)

Jacob Gildenblat, Ido Ben-Shaul, Zvi Lapp, Eldad Klaiman

摘要:

多实例学习是弱监督学习的一种形式,其中数据被安排在称为袋子的实例集合中,每个袋子分配一个标签。通过在实例预测或嵌入上应用置换不变池化算子,从多个实例导出包级类预测。我们提出了一种新的池运营商称为\textbf{Certainty Pooling},它将模型的确定性到袋预测,从而产生一个更强大的和可解释的模型。我们比较我们提出的方法与其他池运营商在控制实验中的低证据比袋MNIST的基础上,以及在现实生活中的组织病理学数据集-Camelyon 16。我们的方法在包级和实例级预测方面都优于其他方法,特别是当只有小的训练集可用时。我们讨论了我们的方法背后的基本原理及其在这些类型的数据集上的优越性。

英文摘要:

Multiple Instance Learning is a form of weakly supervised learning in which the data is arranged in sets of instances called bags with one label assigned per bag. The bag level class prediction is derived from the multiple instances through application of a permutation invariant pooling operator on instance predictions or embeddings. We present a novel pooling operator called \textbf{Certainty Pooling} which incorporates the model certainty into bag predictions resulting in a more robust and explainable model. We compare our proposed method with other pooling operators in controlled experiments with low evidence ratio bags based on MNIST, as well as on a real life histopathology dataset - Camelyon16. Our method outperforms other methods in both bag level and instance level prediction, especially when only small training sets are available. We discuss the rationale behind our approach and the reasons for its superiority for these types of datasets.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑