Deep Neural Model Inspection and Comparison via Functional Neuron Pathways
Deep Neural Model Inspection and Comparison via Functional Neuron Pathways
复制标题
DOI：
10.18653/v1/p19-1575
复制
发表时间：
2019-07
期刊：
影响因子：
--
通讯作者：
James Fiacco;Samridhi Choudhary;C. Rosé
中科院分区：
文献类型：
--
作者：
James Fiacco;Samridhi Choudhary;C. Rosé
我们介绍了一种解释和比较神经模型的通用方法。该方法用于将复杂的神经模型分解为其功能组件，这些组件由跨网络架构层的共放电神经元组成，我们称之为神经通路。这些路径的功能可以通过识别相关的任务水平和语言启发式来理解，以这种方式，这些知识作为一个透镜，近似网络已经学会了将其应用于预期任务。作为调查这些路径效用的案例研究，我们提出了在两个标准任务（即命名实体识别和识别文本蕴涵）训练的模型中识别的路径的检查。
We introduce a general method for the interpretation and comparison of neural models. The method is used to factor a complex neural model into its functional components, which are comprised of sets of co-firing neurons that cut across layers of the network architecture, and which we call neural pathways. The function of these pathways can be understood by identifying correlated task level and linguistic heuristics in such a way that this knowledge acts as a lens for approximating what the network has learned to apply to its intended task. As a case study for investigating the utility of these pathways, we present an examination of pathways identified in models trained for two standard tasks, namely Named Entity Recognition and Recognizing Textual Entailment.