Neural Power Units
Neural Power Units
复制标题
神经动力单元
DOI：
--
复制
发表时间：
2020
期刊：
影响因子：
--
通讯作者：
V. Šmídl
中科院分区：
文献类型：
--
作者：
Niklas Heim;T. Pevný;V. Šmídl
传统的神经网络可以近似简单的算术运算，但无法推广到训练过程中看到的数字范围之外。神经算术单元旨在克服这一困难，但目前的算术单元要么仅限于对正数进行运算，要么只能表示算术运算的子集。我们介绍了神经功率单元（NPU），它在真实的数字的全域上操作，并且能够在单层中学习任意功率函数。因此，NPU修复了现有算术单元的缺点，并扩展了它们的表达能力。我们通过使用复数运算来实现这一点，而不需要将网络转换为复数。将单元简化为RealNPU会产生高度可解释的模型。我们表明，NPU在人工算术数据集上的准确性和稀疏性方面优于竞争对手，并且RealNPU只能从数据中发现动力系统的控制方程。
Conventional Neural Networks can approximate simple arithmetic operations, but fail to generalize beyond the range of numbers that were seen during training. Neural Arithmetic Units aim to overcome this difficulty, but current arithmetic units are either limited to operate on positive numbers or can only represent a subset of arithmetic operations. We introduce the Neural Power Unit (NPU) that operates on the full domain of real numbers and is capable of learning arbitrary power functions in a single layer. The NPU thus fixes the shortcomings of existing arithmetic units and extends their expressivity. We achieve this by using complex arithmetic without requiring a conversion of the network to complex numbers. A simplification of the unit to the RealNPU yields a highly interpretable model. We show that the NPUs outperform their competitors in terms of accuracy and sparsity on artificial arithmetic datasets, and that the RealNPU can discover the governing equations of a dynamical systems only from data.
影响因子：
4.6
作者：
Taghvaei A;Georgiou TT;Norton L;Tannenbaum A
通讯作者：
Tannenbaum A
可解释的多项式神经常微分方程。
DOI：
10.1063/5.0130803
发表时间：
2023
期刊：
Chaos (Woodbury, N.Y.)
影响因子：
--
作者：
Fronk,Colby;Petzold,Linda
通讯作者：
Petzold,Linda