arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2006.01681cs.LGcs.NEstat.ML
收藏

神经动力单元

Neural Power Units

Artificial Intelligence Center(人工智能中心);Czech Technical University(捷克技术大学)

Niklas Heim, Tomáš Pevný, Václav Šmídl

摘要:

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),它在实数的整个域上运行,并且能够在单层中学习任意的幂函数。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会产生高度透明的模型。我们表明,NPU在人工算术数据集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。

英文摘要:

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 transparent 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 system only from data.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑