BriefGPT.xyz
Jan, 2022
多实例学习的模型无关可解释性
Model Agnostic Interpretability for Multiple Instance Learning
HTML
PDF
Joseph Early, Christine Evers, Sarvapali Ramchurn
TL;DR
本文研究了多示例学习（MIL）中模型的可解释性，并提出了几种模型无关的方法来满足这些要求，在多个数据集上与现有的基于模型的MIL模型进行比较，并取得了高达30%的解释性准确性的提高。同时研究了这些方法识别实例间相互作用的能力和扩展到大型数据集，从而提高了它们应对实际问题的能力。
Abstract
In
multiple instance learning
(MIL),
models
are trained using bags of instances, where only a single label is provided for each bag. A bag label is often only determined by a handful of key instances within a bag
→