Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

The backpropagation algorithm implemented on spiking neuromorphic hardware

Journal Article · · Nature Communications
 [1];  [2];  [3];  [4];  [3]
  1. University of Zurich and ETH Zurich (Switzerland); Forschungszentrum Jülich (Germany)
  2. Los Alamos National Laboratory (LANL), Los Alamos, NM (United States); London Institute for Mathematical Sciences (United Kingdom)
  3. Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)
  4. Peking University, Beijing (China)
The capabilities of natural neural systems have inspired both new generations of machine learning algorithms as well as neuromorphic, very large-scale integrated circuits capable of fast, low-power information processing. However, it has been argued that most modern machine learning algorithms are not neurophysiologically plausible. In particular, the workhorse of modern deep learning, the backpropagation algorithm, has proven difficult to translate to neuromorphic hardware. This study presents a neuromorphic, spiking backpropagation algorithm based on synfire-gated dynamical information coordination and processing implemented on Intel’s Loihi neuromorphic research processor. We demonstrate a proof-of-principle three-layer circuit that learns to classify digits and clothing items from the MNIST and Fashion MNIST datasets. To our knowledge, this is the first work to show a Spiking Neural Network implementation of the exact backpropagation algorithm that is fully on-chip without a computer in the loop. It is competitive in accuracy with off-chip trained SNNs and achieves an energy-delay product suitable for edge computing. This implementation shows a path for using in-memory, massively parallel neuromorphic processors for low-power, low-latency implementation of modern deep learning applications.
Research Organization:
Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)
Sponsoring Organization:
Natural Science Foundation of China; Swiss National Science Foundation (SNSF); USDOE Laboratory Directed Research and Development (LDRD) Program; USDOE National Nuclear Security Administration (NNSA); University of Zurich
Grant/Contract Number:
89233218CNA000001
OSTI ID:
2476747
Alternate ID(s):
OSTI ID: 2532373
Report Number(s):
LA-UR--21-24457
Journal Information:
Nature Communications, Journal Name: Nature Communications Journal Issue: 1 Vol. 15; ISSN 2041-1723
Publisher:
Nature Publishing GroupCopyright Statement
Country of Publication:
United States
Language:
English

References (87)

Difference Target Propagation book January 2015
Taylor expansion of the accumulated rounding error journal June 1976
A mechanism for graded, dynamically routable current propagation in pulse-gated synfire chains and implications for information coding journal August 2015
NengoDL: Combining Deep Learning and Neuromorphic Modelling Methods journal April 2019
Error-backpropagation in temporally encoded networks of spiking neurons journal October 2002
Dendritic solutions to the credit assignment problem journal February 2019
Deep learning in spiking neural networks journal March 2019
Recurrent Network Models of Sequence Generation and Memory journal April 2016
The recent excitement about neural networks journal January 1989
An ultra-sparse code underliesthe generation of neural sequences in a songbird journal September 2002
Reverse replay of behavioural sequences in hippocampal place cells during the awake state journal February 2006
Random synaptic feedback weights support error backpropagation for deep learning journal November 2016
Using goal-driven deep learning models to understand sensory cortex journal February 2016
Control of synaptic plasticity in deep cortical networks journal February 2018
Packet-based communication in the cortex journal October 2015
A solution to the learning dilemma for recurrent networks of spiking neurons journal July 2020
Backpropagation and the brain journal April 2020
Burst-dependent synaptic plasticity can coordinate learning in hierarchical circuits journal May 2021
Experimental Demonstration of Supervised Learning in Spiking Neural Networks with Phase-Change Memory Synapses journal May 2020
Training deep neural networks for binary communication with the Whetstone method journal January 2019
Rapid online learning and robust recall in a neuromorphic olfactory circuit journal March 2020
Optimized spiking neurons can classify images with high accuracy through temporal coding with two spikes journal March 2021
Fast and energy-efficient neuromorphic deep learning with first-spike times journal September 2021
Synthesizing cognition in neuromorphic electronic systems journal July 2013
Convolutional networks for fast, energy-efficient neuromorphic computing journal September 2016
Simple framework for constructing functional spiking recurrent neural networks journal October 2019
Solving the Distal Reward Problem through Linkage of STDP and Dopamine Signaling journal January 2007
Responses of neurons in primary and inferior temporal visual cortices to natural scenes
  • Baddeley, Roland; Abbott, L. F.; Booth, Michael C. A.
  • Proceedings of the Royal Society of London. Series B: Biological Sciences, Vol. 264, Issue 1389 https://doi.org/10.1098/rspb.1997.0246
journal December 1997
Neuromorphic electronic systems journal January 1990
Gradient-based learning applied to document recognition journal January 1998
A pulse-gated, predictive neural circuit conference November 2016
On-chip Few-shot Learning with Surrogate Gradient Descent on a Neuromorphic Processor conference August 2020
SpiNNaker: A multi-core System-on-Chip for massively-parallel neural net simulation conference September 2012
In-Hardware Learning of Multilayer Spiking Neural Networks on a Neuromorphic Processor conference December 2021
Temporal Coding in Spiking Neural Networks with Alpha Synaptic Function conference May 2020
Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification conference December 2015
Backpropagation without weight transport conference June 1994
Is backpropagation biologically plausible? conference January 1989
Scalable energy-efficient, low-latency implementations of trained spiking Deep Belief Networks on SpiNNaker conference July 2015
A wafer-scale neuromorphic hardware system for large-scale neural modeling
  • Schemmel, Johannes; Briiderle, Daniel; Griibl, Andreas
  • 2010 IEEE International Symposium on Circuits and Systems - ISCAS 2010, Proceedings of 2010 IEEE International Symposium on Circuits and Systems https://doi.org/10.1109/ISCAS.2010.5536970
conference May 2010
Conversion of analog to spiking neural networks using sparse temporal coding conference January 2018
A 28-nm Convolutional Neuromorphic Processor Enabling Online Learning with Spike-Based Retinas conference October 2020
Visual Pattern Recognition with on On-Chip Learning: Towards a Fully Neuromorphic Approach conference October 2020
7.6 A 65nm 236.5nJ/Classification Neuromorphic Processor with 7.5% Energy Overhead On-Chip Learning Using Direct Spike-Only Feedback conference February 2019
On-Chip Error-Triggered Learning of Multi-Layer Memristive Spiking Neural Networks journal December 2020
Advancing Neuromorphic Computing With Loihi: A Survey of Results and Outlook journal May 2021
A 4096-Neuron 1M-Synapse 3.8-pJ/SOP Spiking Neural Network With On-Chip STDP Learning and Sparse Weights in 10-nm FinFET CMOS journal April 2019
Programming Spiking Neural Networks on Intel’s Loihi journal March 2018
Loihi: A Neuromorphic Manycore Processor with On-Chip Learning journal January 2018
Surrogate Gradient Learning in Spiking Neural Networks: Bringing the Power of Gradient-Based Optimization to Spiking Neural Networks journal November 2019
Efficient Neuromorphic Signal Processing with Loihi 2 conference October 2021
Demonstrating Hybrid Learning in a Flexible Neuromorphic Hardware System journal February 2017
Toward High-Accuracy and Low-Latency Spiking Neural Networks With Two-Stage Optimization journal January 2024
A 640M pixel/s 3.65mW sparse event-driven neuromorphic object recognition processor with on-chip learning conference June 2015
Competitive Learning: From Interactive Activation to Adaptive Resonance journal January 1987
Synfire Chains and Cortical Songs: Temporal Modules of Cortical Activity journal April 2004
Spike Synchronization and Rate Modulation Differentially Involved in Motor Cortical Function journal December 1997
Awake hippocampal reactivations project onto orthogonal neuronal assemblies journal September 2016
A Pulse-gated, Neural Implementation of the Backpropagation Algorithm conference March 2019
Efficient Biologically-Plausible Training of Spiking Neural Networks with Precise Timing conference July 2021
NxTF: An API and Compiler for Deep Spiking Neural Networks on Intel Loihi
  • Rueckauer, Bodo; Bybee, Connor; Goettsche, Ralf
  • ACM Journal on Emerging Technologies in Computing Systems, Vol. 18, Issue 3 https://doi.org/10.1145/3501770
journal July 2022
Spatiotemporal firing patterns in the frontal cortex of behaving monkeys journal October 1993
Neuromorphic Backpropagation Algorithm Software software April 2022
Learning Only When Necessary: Better Memories of Correlated Patterns in Networks with Bounded Synapses journal October 2005
Supervised Learning in Multilayer Spiking Neural Networks journal February 2013
Biologically Plausible Error-Driven Learning Using Local Activation Differences: The Generalized Recirculation Algorithm journal July 1996
Optimal Spike-Timing-Dependent Plasticity for Precise Action Potential Firing in Supervised Learning journal June 2006
SuperSpike: Supervised Learning in Multilayer Spiking Neural Networks journal June 2018
A Mechanism for Synaptic Copy Between Neural Circuits journal October 2019
A Learning Theory for Reward-Modulated Spike-Timing-Dependent Plasticity with Application to Biofeedback journal October 2008
Graded, Dynamically Routable Information Processing with Synfire-Gated Synfire Chains journal June 2016
How Important Is Weight Symmetry in Backpropagation? journal February 2016
A 3.43TOPS/W 48.9pJ/pixel 50.1nJ/classification 512 analog neuron sparse coding neural network with on-chip learning and classification in 40nm CMOS conference June 2017
Nengo and Low-Power AI Hardware for Robust, Embedded Neurorobotics journal October 2020
A reconfigurable on-line learning spiking neuromorphic processor comprising 256 neurons and 128K synapses journal April 2015
Conversion of Continuous-Valued Deep Networks to Efficient Event-Driven Networks for Image Classification journal December 2017
Spatio-Temporal Backpropagation for Training High-Performance Spiking Neural Networks journal May 2018
Going Deeper in Spiking Neural Networks: VGG and Residual Architectures journal March 2019
Mixed-Precision Deep Learning Based on Computational Memory journal May 2020
Mutual Information and Information Gating in Synfire Chains journal February 2018
A Review of Binarized Neural Networks journal June 2019
A Fokker-Planck approach to graded information propagation in pulse-gated feedforward neuronal networks preprint January 2015
Gradient Descent for Spiking Neural Networks preprint January 2017
SLAYER: Spike Layer Error Reassignment in Time preprint January 2018
Dendritic cortical microcircuits approximate the backpropagation algorithm preprint January 2018
Event-based Backpropagation for Analog Neuromorphic Hardware preprint January 2023
Fast and flexible sequence induction in spiking neural networks via rapid excitability changes journal May 2019

Similar Records

Neuromorphic Backpropagation Algorithm Software
Software · Wed Apr 13 20:00:00 EDT 2022 · OSTI ID:code-74250

Solving sparse finite element problems on neuromorphic hardware
Journal Article · Wed Nov 12 19:00:00 EST 2025 · Nature Machine Intelligence · OSTI ID:3006560