Multiple instance learning (MIL) assigns a single class label to a bag of instances tailored for some real-world applications such as drug activity prediction. Classical MIL methods focus on figuring out interested instances, that is, region of interests (ROIs). However, owing to the non-differentiable selection process, these methods are not feasible in deep learning. Thus, we focus on fusing ROIs identification with deep MILs in this paper. We propose a novel deep MIL framework based on hard selection, that is, deep multiple instance selection (DMIS), which can automatically figure ROIs out in an end-to-end approach. To be specific, we propose DMIS-GS for instance selection via gumbel softmax or gumbel top-k, and then make predictions for this bag without the interference of redundant instances. For balancing exploration and exploitation of key instances, we apply a cooling down approach to the temperature in DMIS-GS, and propose a variance normalization method to make this hyper-parameter tuning process much easier. Generally, we give a theoretical analysis of our framework. The empirical investigations reveal the proposed frameworks’ superiorities against classical MIL methods on generalization ability, positioning ROIs, and comprehensibility on both synthetic and real-world datasets.

多实例学习（MIL）将单个类别标签分配给针对某些实际应用（例如毒品活动预测）量身定制的实例包。传统的MIL方法专注于找出感兴趣的实例，即感兴趣区域（ROI）。然而，由于不可微的选择过程，这些方法在深度学习中不可行。因此，在本文中，我们着重将ROI识别与深层MIL融合在一起。我们提出了一种基于硬选择的新型深度MIL框架，即深度多实例选择（DMIS），它可以以端到端的方式自动计算出ROI。具体来说，我们建议使用DMIS-GS，例如通过gumbel softmax或gumbel top- k选择，然后在没有冗余实例干扰的情况下对此包进行预测。为了平衡关键实例的探索和开发，我们对DMIS-GS中的温度采用了冷却方法，并提出了方差归一化方法，以使此超参数调整过程变得更加容易。通常，我们会对我们的框架进行理论分析。实证研究表明，在综合和真实数据集上，提出的框架相对于经典MIL方法在泛化能力，定位ROI和可理解性方面具有优势。