Abernot Madeleine, Azemard Nadine, Todri-Sanial Aida
Laboratoire d'Informatique de Robotique et de Microélectronique de Montpellier (LIRMM), Department of Microelectroncis, University of Montpellier, CNRS, Montpellier, France.
Electrical Engineering Department, Eindhoven University of Technology, Eindhoven, Netherlands.
Front Neurosci. 2023 Jun 15;17:1196796. doi: 10.3389/fnins.2023.1196796. eCollection 2023.
In the human brain, learning is continuous, while currently in AI, learning algorithms are pre-trained, making the model non-evolutive and predetermined. However, even in AI models, environment and input data change over time. Thus, there is a need to study continual learning algorithms. In particular, there is a need to investigate how to implement such continual learning algorithms on-chip. In this work, we focus on Oscillatory Neural Networks (ONNs), a neuromorphic computing paradigm performing auto-associative memory tasks, like Hopfield Neural Networks (HNNs). We study the adaptability of the HNN unsupervised learning rules to on-chip learning with ONN. In addition, we propose a first solution to implement unsupervised on-chip learning using a digital ONN design. We show that the architecture enables efficient ONN on-chip learning with Hebbian and Storkey learning rules in hundreds of microseconds for networks with up to 35 fully-connected digital oscillators.
在人类大脑中，学习是持续不断的，而目前在人工智能领域，学习算法是经过预训练的，这使得模型不可进化且具有预定性。然而，即使在人工智能模型中，环境和输入数据也会随时间变化。因此，有必要研究持续学习算法。特别是，有必要研究如何在芯片上实现这种持续学习算法。在这项工作中，我们专注于振荡神经网络（ONN），它是一种执行自联想记忆任务的神经形态计算范式，类似于霍普菲尔德神经网络（HNN）。我们研究了HNN无监督学习规则对使用ONN进行片上学习的适应性。此外，我们提出了一种使用数字ONN设计实现无监督片上学习的初步解决方案。我们表明，该架构能够在数百微秒内，通过Hebbian和Storkey学习规则，为具有多达35个全连接数字振荡器的网络实现高效的ONN片上学习。
