[论文解读] Understanding Community Structure in Layered Neural Networks
本文提出一种定量方法,通过定义输入到社区和社区到输出的影响,解释分层神经网络中社区的作用,实现对模块化结构的精确分析。该方法揭示了每个社区在推理中的贡献,克服了依赖阈值的社区检测方法的局限性,并在不同超参数下提供了稳定且可解释的洞察。
A layered neural network is now one of the most common choices for the prediction of high-dimensional practical data sets, where the relationship between input and output data is complex and cannot be represented well by simple conventional models. Its effectiveness is shown in various tasks, however, the lack of interpretability of the trained result by a layered neural network has limited its application area. In our previous studies, we proposed methods for extracting a simplified global structure of a trained layered neural network by classifying the units into communities according to their connection patterns with adjacent layers. These methods provided us with knowledge about the strength of the relationship between communities from the existence of bundled connections, which are determined by threshold processing of the connection ratio between pairs of communities. However, it has been difficult to understand the role of each community quantitatively by observing the modular structure. We could only know to which sets of the input and output dimensions each community was mainly connected, by tracing the bundled connections from the community to the input and output layers. Another problem is that the finally obtained modular structure is changed greatly depending on the setting of the threshold hyperparameter used for determining bundled connections. In this paper, we propose a new method for interpreting quantitatively the role of each community in inference, by defining the effect of each input dimension on a community, and the effect of a community on each output dimension. We show experimentally that our proposed method can reveal the role of each part of a layered neural network by applying the neural networks to three types of data sets, extracting communities from the trained network, and applying the proposed method to the community structure.
研究动机与目标
- 为解决分层神经网络(LNNs)中可解释性不足的问题,特别是理解检测到的社区的功能角色。
- 克服基于阈值的社区检测方法带来的不稳定性与主观性,这些方法会因超参数选择而改变模块化结构。
- 开发一种定量框架,用于衡量输入维度对社区的影响以及社区对输出维度的影响。
- 通过用基于影响的社区角色量化替代启发式捆绑连接阈值,实现在分层神经网络中一致且可解释的分析。
提出的方法
- 引入输入影响和输出影响度量,以量化每个输入维度对社区的影响以及每个社区对输出维度的影响。
- 使用连接权重的加权和定义社区层面的影响:$\tau^{\prime+}_{c,j} = \frac{\sum_k B^{+}_{k,j} q_{k,c}}{\sum_k q_{c,k}}$ 和 $\tau^{\prime-}_{c,j} = \frac{\sum_k B^{-}_{k,j} q_{k,c}}{\sum_k q_{k,c}}$,其中 $q_{k,c}$ 表示社区成员资格。
- 应用期望最大化(EM)算法并结合社区检测试验,以最大化期望对数似然,从而选择最优社区结构。
- 使用归一化数据和LASSO正则化以稳定训练并提升泛化能力。
- 在实验中采用固定训练顺序,以确保社区检测过程中的稳定收敛。
- 在三个数据集上验证该方法:合成数据、消费者价格指数数据以及图像分类数据,并与线性模型进行结果比较。
实验结果
研究问题
- RQ1如何在超越简单连接模式的基础上,对分层神经网络中每个社区的功能角色进行定量测量?
- RQ2所提出的基于影响的方法是否相较于基于阈值的捆绑连接分析,能产生更稳定且可解释的社区解读?
- RQ3该方法能否揭示有意义的输入-输出关系,真实反映社区对推理的实际贡献?
- RQ4在捕捉高维数据中复杂输入-输出关系方面,该方法与线性模型相比表现如何?
- RQ5当基于阈值的方法失效时,该方法的社区结构解读在不同超参数设置下是否依然稳健?
主要发现
- 所提出的方法成功量化了每个输入维度对社区的影响以及每个社区对输出维度的影响,实现了对社区角色的精确解读。
- 该方法在不同超参数下提供了稳定的社区解读,消除了基于阈值方法中常见的不稳定性。
- 在消费者价格指数数据集中,分层神经网络的泛化误差低于线性模型,表明其在处理复杂高维数据方面表现更优。
- 线性模型在输入数据为29个月时达到最小泛化误差,但分层神经网络在所有测试的时间跨度内均表现更优。
- 该方法表明,社区层面的影响比仅依赖捆绑连接更具信息量,为理解推理过程中功能角色提供了更清晰的视角。
更好的研究,从现在开始
从阅读论文到最终审阅,大幅缩短您的研究时间。
无需绑定信用卡
本解读由 AI 生成,并经人工编辑审核。