Loihi: A Neuromorphic Manycore Processor with On-Chip Learning

被引:2390
作者
Davies, Mike [1 ]
Srinivasa, Narayan [2 ]
Lin, Tsung-Han [1 ]
Chinya, Gautham [1 ]
Cao, Yongqiang [1 ]
Choday, Sri Harsha [1 ]
Dimou, Georgios [3 ]
Joshi, Prasad [1 ]
Imam, Nabil [1 ]
Jain, Shweta [1 ]
Liao, Yuyun [1 ]
Lin, Chit-Kwan [1 ]
Lines, Andrew [1 ]
Liu, Ruokun [1 ]
Mathaikutty, Deepak [1 ]
Mccoy, Steve [1 ]
Paul, Arnab [1 ]
Tse, Jonathan [1 ]
Venkataramanan, Guruguhanathan [1 ]
Weng, Yi-Hsin [1 ]
Wild, Andreas [1 ]
Yang, Yoonseok [1 ]
Wang, Hong [1 ]
机构
[1] Intel Corp, Intel Labs, Santa Clara, CA 95054 USA
[2] Eta Compute, Westlake Village, CA 91362 USA
[3] Reduced Energy Microsyst, San Francisco, CA USA
关键词
artificial intelligence; machine learning; neuromorphic computing;
D O I
10.1109/MM.2018.112130359
中图分类号
TP3 [计算技术、计算机技术];
学科分类号
0812 ;
摘要
Loihi is a 60-mm(2) chip fabricated in Intel's 14-nm process that advances the state-of-the-art modeling of spiking neural networks in silicon. It integrates a wide range of novel features for the field, such as hierarchical connectivity, dendritic compartments, synaptic delays, and, most importantly, programmable synaptic learning rules. Running a spiking convolutional form of the Locally Competitive Algorithm, Loihi can solve LASSO optimization problems with over three orders of magnitude superior energy-delay product compared to conventional solvers running on a CPU isoprocess/voltage/area. This provides an unambiguous example of spike-based computation, outperforming all known conventional solutions.
引用
收藏
页码:82 / 99
页数:18
相关论文
共 14 条
[1]   A Fast Iterative Shrinkage-Thresholding Algorithm for Linear Inverse Problems [J].
Beck, Amir ;
Teboulle, Marc .
SIAM JOURNAL ON IMAGING SCIENCES, 2009, 2 (01) :183-202
[2]   Neural Dynamics as Sampling: A Model for Stochastic Computation in Recurrent Networks of Spiking Neurons [J].
Buesing, Lars ;
Bill, Johannes ;
Nessler, Bernhard ;
Maass, Wolfgang .
PLOS COMPUTATIONAL BIOLOGY, 2011, 7 (11)
[3]   Reinforcement learning through modulation of spike-timing-dependent synaptic plasticity [J].
Florian, Razvan V. .
NEURAL COMPUTATION, 2007, 19 (06) :1468-1502
[4]   A triplet spike-timing-dependent plasticity model generalizes the Bienenstock-Cooper-Munro rule to higher-order spatiotemporal correlations [J].
Gjorgjieva, Julijana ;
Clopath, Claudia ;
Audet, Juliette ;
Pfister, Jean-Pascal .
PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA, 2011, 108 (48) :19383-19388
[5]   Polychronization: Computation with spikes [J].
Izhikevich, Eugene M. .
NEURAL COMPUTATION, 2006, 18 (02) :245-282
[6]  
Lin T. - H., 2017, LOCAL INFORM FEEDBAC
[7]   A million spiking-neuron integrated circuit with a scalable communication network and interface [J].
Merolla, Paul A. ;
Arthur, John V. ;
Alvarez-Icaza, Rodrigo ;
Cassidy, Andrew S. ;
Sawada, Jun ;
Akopyan, Filipp ;
Jackson, Bryan L. ;
Imam, Nabil ;
Guo, Chen ;
Nakamura, Yutaka ;
Brezzo, Bernard ;
Vo, Ivan ;
Esser, Steven K. ;
Appuswamy, Rathinakumar ;
Taba, Brian ;
Amir, Arnon ;
Flickner, Myron D. ;
Risk, William P. ;
Manohar, Rajit ;
Modha, Dharmendra S. .
SCIENCE, 2014, 345 (6197) :668-673
[8]   Event-Driven Random Back-Propagation: Enabling Neuromorphic Deep Learning Machines [J].
Neftci, Emre O. ;
Augustine, Charles ;
Paul, Somnath ;
Detorakis, Georgios .
FRONTIERS IN NEUROSCIENCE, 2017, 11
[9]   Rapid, parallel path planning by propagating wavefronts of spiking neural activity [J].
Ponulak, Filip ;
Hopfield, John J. .
FRONTIERS IN COMPUTATIONAL NEUROSCIENCE, 2013, 7
[10]   Supervised Learning in Spiking Neural Networks with ReSuMe: Sequence Learning, Classification, and Spike Shifting [J].
Ponulak, Filip ;
Kasinski, Andrzej .
NEURAL COMPUTATION, 2010, 22 (02) :467-510