Boolean learning under noise-perturbations in hardware neural networks

Nanophotonics

article

Louis Andreoli¹  Xavier Porte¹  Stéphane Chrétien¹  Maxime Jacquot¹  Laurent Larger¹  Daniel Brunner¹ 

[1] Univ. Bourgogne Franche-Comté;Laboratoire ERIC, France National Physical Laboratory, UK The Alan Turing Institute

关键词: Boolean learning;    neural networks;    noise;   

DOI  :  10.1515/nanoph-2020-0171

学科分类：社会科学、人文和艺术（综合）

来源: De Gruyter

【 摘 要 】

A high efficiency hardware integration of neural networks benefits from realizing nonlinearity, network connectivity and learning fully in a physical substrate. Multiple systems have recently implemented some or all of these operations, yet the focus was placed on addressing technological challenges. Fundamental questions regarding learning in hardware neural networks remain largely unexplored. Noise in particular is unavoidable in such architectures, and here we experimentally and theoretically investigate its interaction with a learning algorithm using an opto-electronic recurrent neural network. We find that noise strongly modifies the system’s path during convergence, and surprisingly fully decorrelates the final readout weight matrices. This highlights the importance of understanding architecture, noise and learning algorithm as interacting players, and therefore identifies the need for mathematical tools for noisy, analogue system optimization.

【 授权许可】

CC BY