Neural network interpretation using descrambler groups.
Neural network interpretation using descrambler groups.
复制标题
DOI：
10.1073/pnas.2016917118
复制
发表时间：
2021-02-02
影响因子：
11.1
通讯作者：
Kuprov I
中科院分区：
文献类型：
--
作者：
Amey JL;Keeley J;Choudhury T;Kuprov I
人工神经网络是出了名的不透明，人们常常不清楚它们是如何工作的。在这篇文章中，我们提出了一种群体理论的方法来找出答案。它揭示了相当多的内部复杂性，即使在简单的神经网络中：我们的网络显然发明了一个优雅的数字滤波器，正则化积分变换，甚至切比雪夫多项式。这是拯救还原论的一步。几个世纪以来，科学的哲学方法一直是寻找支配现实的基本定律，测试这些定律，并利用它们的预测能力。黑盒神经网络在这所学校里相当于亵渎神明，但它们是不可抗拒的，因为它们“只是工作”。解释它们是如何工作的是一个众所周知的难题，本文提供了部分解决方案。缺乏可解释性和信任是深度神经网络备受批评的一个特征。在全连接网络中，内层之间的信令被加扰，因为反向传播训练不需要以任何特定顺序排列感知器。结果是一个黑盒子;这个问题在科学计算和数字信号处理（DSP）中尤其严重，其中神经网络执行不简化为特征或概念的抽象数学变换。我们在这里提出了一个组理论的过程，试图把内层信令成一个人类可读的形式，假设这种形式存在，并具有可识别和可量化的功能，例如，平滑度或局部性。我们将该方法应用于DEERNet（电子自旋共振中使用的DSP网络）并设法解扰它。我们发现了相当大的内部复杂性：该网络自发地发明了带通滤波器、陷波滤波器、频率轴重标度变换、频分复用、群嵌入、频谱滤波正则化以及从调和函数到切比雪夫多项式的映射-在10分钟的无人值守训练中从随机初始猜测。
Artificial neural networks are famously opaque—it is often unclear how they work. In this communication, we propose a group-theoretical way of finding out. It reveals considerable internal sophistication, even in simple neural networks: our nets apparently invented an elegant digital filter, a regularized integral transform, and even Chebyshev polynomials. This is a step toward saving reductionism. For centuries, the philosophical approach to science has been to find fundamental laws that govern reality, to test those laws, and to use their predictive power. Black-box neural networks amount to blasphemy within that school, but they are irresistible because they “just work.” Explaining how they work is a notoriously difficult problem, to which this paper offers a partial solution. The lack of interpretability and trust is a much-criticized feature of deep neural networks. In fully connected nets, the signaling between inner layers is scrambled because backpropagation training does not require perceptrons to be arranged in any particular order. The result is a black box; this problem is particularly severe in scientific computing and digital signal processing (DSP), where neural nets perform abstract mathematical transformations that do not reduce to features or concepts. We present here a group-theoretical procedure that attempts to bring inner-layer signaling into a human-readable form, the assumption being that this form exists and has identifiable and quantifiable features—for example, smoothness or locality. We applied the proposed method to DEERNet (a DSP network used in electron spin resonance) and managed to descramble it. We found considerable internal sophistication: the network spontaneously invents a bandpass filter, a notch filter, a frequency axis rescaling transformation, frequency-division multiplexing, group embedding, spectral filtering regularization, and a map from harmonic functions into Chebyshev polynomials—in 10 min of unattended training from a random initial guess.
登录
查看更多内容
影响因子：
6
作者：
Effati, Sohrab;Buzhabadi, Reza
通讯作者：
Buzhabadi, Reza
影响因子：
64.8
作者：
RUMELHART, DE;HINTON, GE;WILLIAMS, RJ
通讯作者：
WILLIAMS, RJ
影响因子：
2.7
作者：
LIU, DC;NOCEDAL, J
通讯作者：
NOCEDAL, J
影响因子：
22.7
作者：
Lipton, Zachary C.
通讯作者：
Lipton, Zachary C.
影响因子：
4
作者：
Gnecco, Giorgio;Kurkova, Vera;Sanguineti, Marcello
通讯作者：
Sanguineti, Marcello