PDF(Sci-hub)
Abstract：
Deep neural networks (DNNs) can resolve real-world categorization tasks with apparent human-level performance. However, true equivalence of behavioral performance between humans and their DNN models requires that their internal mechanisms process equivalent features of the stimulus. To develop such feature equivalence, our methodology leveraged an interpretable and experimentally controlled generative model of the stimuli (realistic three-dimensional textured faces). Humans rated the similarity of randomly generated faces to four familiar identities. We predicted these similarity ratings from the activations of five DNNs trained with different optimization objectives. Using information theoretic redundancy, reverse correlation, and the testing of generalization gradients, we show that DNN predictions of human behavior improve because their shape and texture features overlap with those that subsume human behavior. Thus, we must equate the functional features that subsume the behavioral performances of the brain and its models before comparing where, when, and how these features are processed.
摘要：
深度神经网络（DNN）可以解决现实世界的分类任务，具有明显的人类水平的性能。然而,人类与其DNN模型之间的行为表现的真正等效性要求其内部机制处理刺激的等效特征。为了开发这种特征等效性，我们的方法利用了可解释和实验控制的刺激生成模型（逼真的三维纹理面）。人类对随机生成的面孔与四个熟悉的身份的相似性进行了评分。我们从使用不同优化目标训练的五个DNN的激活中预测了这些相似性评级。利用信息理论冗余，反向相关，以及泛化梯度的测试，我们表明，DNN对人类行为的预测有所改善，因为它们的形状和纹理特征与包含人类行为的特征重叠。因此，在比较之前，我们必须将包含大脑行为表现的功能特征及其模型等同起来，when,以及如何处理这些特征。
