Neural and Synaptic Array Transceiver: A Brain-Inspired Computing Framework for Embedded Learning.
Neural and Synaptic Array Transceiver: A Brain-Inspired Computing Framework for Embedded Learning.
复制标题
DOI：
10.3389/fnins.2018.00583
复制
发表时间：
2018
影响因子：
4.3
通讯作者：
Neftci E
中科院分区：
文献类型：
--
作者：
Detorakis G;Sheik S;Augustine C;Paul S;Pedroni BU;Dutt N;Krichmar J;Cauwenberghs G;Neftci E
自主和自适应行为的嵌入式持续学习是神经形态硬件的关键应用。然而，由于缺乏合适的算法框架，灵活高效的大规模嵌入式学习的神经形态实现受到了阻碍。因此，大多数神经形态硬件在大型专用处理器或gpu集群上进行离线训练，并在事后转移到设备上。我们通过引入神经和突触阵列收发器（NSAT）来解决这个问题，NSAT是一种神经形态计算框架，通过匹配算法要求和神经和突触动力学来促进灵活高效的嵌入式学习。NSAT支持事件驱动的有监督、无监督和强化学习算法，包括深度学习。我们在广泛的任务中展示了NSAT，包括Mihalas-Niebur神经元的模拟、动态神经场、基于事件的深度学习的事件驱动随机反向传播、用于无监督学习的基于事件的对比发散以及用于序列学习的基于电压的学习规则。我们预计，这一贡献将为新一代设备奠定基础，使自适应移动系统、可穿戴设备和具有数据驱动自主性的机器人成为可能。
Embedded, continual learning for autonomous and adaptive behavior is a key application of neuromorphic hardware. However, neuromorphic implementations of embedded learning at large scales that are both flexible and efficient have been hindered by a lack of a suitable algorithmic framework. As a result, most neuromorphic hardware are trained off-line on large clusters of dedicated processors or GPUs and transferred post hoc to the device. We address this by introducing the neural and synaptic array transceiver (NSAT), a neuromorphic computational framework facilitating flexible and efficient embedded learning by matching algorithmic requirements and neural and synaptic dynamics. NSAT supports event-driven supervised, unsupervised and reinforcement learning algorithms including deep learning. We demonstrate the NSAT in a wide range of tasks, including the simulation of Mihalas-Niebur neuron, dynamic neural fields, event-driven random back-propagation for event-based deep learning, event-based contrastive divergence for unsupervised learning, and voltage-based learning rules for sequence learning. We anticipate that this contribution will establish the foundation for a new generation of devices enabling adaptive mobile systems, wearable devices, and robots with data-driven autonomy.
登录
查看更多内容
影响因子：
25
作者：
Clopath, Claudia;Buesing, Lars;Gerstner, Wulfram
通讯作者：
Gerstner, Wulfram
影响因子：
1.9
作者：
Cerda, Mauricio;Girau, Bernard
通讯作者：
Girau, Bernard
影响因子：
5.4
作者：
Brandli, Christian;Berner, Raphael;Delbruck, Tobi
通讯作者：
Delbruck, Tobi
影响因子：
5.3
作者：
Bi, GQ;Poo, MM
通讯作者：
Poo, MM
影响因子：
1.9
作者：
AMARI, SI
通讯作者：
AMARI, SI