从组合神经网络中获得可靠的解释
Obtaining Faithful Interpretations from Compositional Neural Networks
摘要：
神经模块网络（NMN）是一种流行的组合性建模方法：当应用于语言和视觉问题时，它们实现了高精度，同时反映了网络架构中问题的组成结构。然而，先前的工作隐含地假设，网络模块的结构，描述了抽象的推理过程，提供了一个忠实的解释模型的推理，也就是说，所有的模块执行其预期的行为。在这项工作中，我们提出并进行系统的评估NMN的中间输出NLVR2和DROP，两个数据集，需要组成多个推理步骤。我们发现，中间输出不同于预期的输出，说明网络结构不提供一个忠实的解释模型的行为。为了弥补这一点，我们用辅助监督训练模型，并为模块架构提出特定的选择，以最小的精度代价产生更好的忠诚度。
英文摘要：
Neural module networks (NMNs) are a popular approach for modeling compositionality: they achieve high accuracy when applied to problems in language and vision, while reflecting the compositional structure of the problem in the network architecture. However, prior work implicitly assumed that the structure of the network modules, describing the abstract reasoning process, provides a faithful explanation of the model's reasoning; that is, that all modules perform their intended behaviour. In this work, we propose and conduct a systematic evaluation of the intermediate outputs of NMNs on NLVR2 and DROP, two datasets which require composing multiple reasoning steps. We find that the intermediate outputs differ from the expected output, illustrating that the network structure does not provide a faithful explanation of model behaviour. To remedy that, we train the model with auxiliary supervision and propose particular choices for module architecture that yield much better faithfulness, at a minimal cost to accuracy.
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。
