1
|
Yang S, Gao T, Wang J, Deng B, Lansdell B, Linares-Barranco B. Efficient Spike-Driven Learning With Dendritic Event-Based Processing. Front Neurosci 2021; 15:601109. [PMID: 33679295 PMCID: PMC7933681 DOI: 10.3389/fnins.2021.601109] [Citation(s) in RCA: 60] [Impact Index Per Article: 20.0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/31/2020] [Accepted: 01/21/2021] [Indexed: 11/22/2022] Open
Abstract
A critical challenge in neuromorphic computing is to present computationally efficient algorithms of learning. When implementing gradient-based learning, error information must be routed through the network, such that each neuron knows its contribution to output, and thus how to adjust its weight. This is known as the credit assignment problem. Exactly implementing a solution like backpropagation involves weight sharing, which requires additional bandwidth and computations in a neuromorphic system. Instead, models of learning from neuroscience can provide inspiration for how to communicate error information efficiently, without weight sharing. Here we present a novel dendritic event-based processing (DEP) algorithm, using a two-compartment leaky integrate-and-fire neuron with partially segregated dendrites that effectively solves the credit assignment problem. In order to optimize the proposed algorithm, a dynamic fixed-point representation method and piecewise linear approximation approach are presented, while the synaptic events are binarized during learning. The presented optimization makes the proposed DEP algorithm very suitable for implementation in digital or mixed-signal neuromorphic hardware. The experimental results show that spiking representations can rapidly learn, achieving high performance by using the proposed DEP algorithm. We find the learning capability is affected by the degree of dendritic segregation, and the form of synaptic feedback connections. This study provides a bridge between the biological learning and neuromorphic learning, and is meaningful for the real-time applications in the field of artificial intelligence.
Collapse
Affiliation(s)
- Shuangming Yang
- School of Electrical and Information Engineering, Tianjin University, Tianjin, China
| | - Tian Gao
- School of Electrical and Information Engineering, Tianjin University, Tianjin, China
| | - Jiang Wang
- School of Electrical and Information Engineering, Tianjin University, Tianjin, China
| | - Bin Deng
- School of Electrical and Information Engineering, Tianjin University, Tianjin, China
| | - Benjamin Lansdell
- Department of Bioengineering, University of Pennsylvania, Philadelphia, PA, United States
| | | |
Collapse
|
2
|
Cramer B, Stöckel D, Kreft M, Wibral M, Schemmel J, Meier K, Priesemann V. Control of criticality and computation in spiking neuromorphic networks with plasticity. Nat Commun 2020; 11:2853. [PMID: 32503982 PMCID: PMC7275091 DOI: 10.1038/s41467-020-16548-3] [Citation(s) in RCA: 36] [Impact Index Per Article: 9.0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/21/2019] [Accepted: 04/23/2020] [Indexed: 11/08/2022] Open
Abstract
The critical state is assumed to be optimal for any computation in recurrent neural networks, because criticality maximizes a number of abstract computational properties. We challenge this assumption by evaluating the performance of a spiking recurrent neural network on a set of tasks of varying complexity at - and away from critical network dynamics. To that end, we developed a plastic spiking network on a neuromorphic chip. We show that the distance to criticality can be easily adapted by changing the input strength, and then demonstrate a clear relation between criticality, task-performance and information-theoretic fingerprint. Whereas the information-theoretic measures all show that network capacity is maximal at criticality, only the complex tasks profit from criticality, whereas simple tasks suffer. Thereby, we challenge the general assumption that criticality would be beneficial for any task, and provide instead an understanding of how the collective network state should be tuned to task requirement.
Collapse
Affiliation(s)
- Benjamin Cramer
- Kirchhoff-Institute for Physics, Heidelberg University, Im Neuenheimer Feld 227, 69120, Heidelberg, Germany.
| | - David Stöckel
- Kirchhoff-Institute for Physics, Heidelberg University, Im Neuenheimer Feld 227, 69120, Heidelberg, Germany
| | - Markus Kreft
- Kirchhoff-Institute for Physics, Heidelberg University, Im Neuenheimer Feld 227, 69120, Heidelberg, Germany
| | - Michael Wibral
- Campus Institute for Dynamics of Biological Networks, Georg-August University, Hermann-Rein-Straße 3, 37075, Göttingen, Germany
| | - Johannes Schemmel
- Kirchhoff-Institute for Physics, Heidelberg University, Im Neuenheimer Feld 227, 69120, Heidelberg, Germany
| | - Karlheinz Meier
- Kirchhoff-Institute for Physics, Heidelberg University, Im Neuenheimer Feld 227, 69120, Heidelberg, Germany
| | - Viola Priesemann
- Max-Planck-Institute for Dynamics and Self-Organization, Am Faßberg 17, 37077, Göttingen, Germany.
- Bernstein Center for Computational Neuroscience, Georg-August University, Am Faßberg 17, 37077, Göttingen, Germany.
- Department of Physics, Georg-August University, Friedrich-Hund-Platz 1, 37077, Göttingen, Germany.
| |
Collapse
|
3
|
Stöckel A, Jenzen C, Thies M, Rückert U. Binary Associative Memories as a Benchmark for Spiking Neuromorphic Hardware. Front Comput Neurosci 2017; 11:71. [PMID: 28878642 PMCID: PMC5572441 DOI: 10.3389/fncom.2017.00071] [Citation(s) in RCA: 3] [Impact Index Per Article: 0.4] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/29/2016] [Accepted: 07/20/2017] [Indexed: 11/14/2022] Open
Abstract
Large-scale neuromorphic hardware platforms, specialized computer systems for energy efficient simulation of spiking neural networks, are being developed around the world, for example as part of the European Human Brain Project (HBP). Due to conceptual differences, a universal performance analysis of these systems in terms of runtime, accuracy and energy efficiency is non-trivial, yet indispensable for further hard- and software development. In this paper we describe a scalable benchmark based on a spiking neural network implementation of the binary neural associative memory. We treat neuromorphic hardware and software simulators as black-boxes and execute exactly the same network description across all devices. Experiments on the HBP platforms under varying configurations of the associative memory show that the presented method allows to test the quality of the neuron model implementation, and to explain significant deviations from the expected reference output.
Collapse
Affiliation(s)
- Andreas Stöckel
- Cognitronics and Sensor Systems, Cluster of Excellence Cognitive Interaction Technology, Faculty of Technology, Bielefeld UniversityBielefeld, Germany
| | - Christoph Jenzen
- Cognitronics and Sensor Systems, Cluster of Excellence Cognitive Interaction Technology, Faculty of Technology, Bielefeld UniversityBielefeld, Germany
| | - Michael Thies
- Cognitronics and Sensor Systems, Cluster of Excellence Cognitive Interaction Technology, Faculty of Technology, Bielefeld UniversityBielefeld, Germany
| | - Ulrich Rückert
- Cognitronics and Sensor Systems, Cluster of Excellence Cognitive Interaction Technology, Faculty of Technology, Bielefeld UniversityBielefeld, Germany
| |
Collapse
|
4
|
Friedmann S, Schemmel J, Grubl A, Hartel A, Hock M, Meier K. Demonstrating Hybrid Learning in a Flexible Neuromorphic Hardware System. IEEE TRANSACTIONS ON BIOMEDICAL CIRCUITS AND SYSTEMS 2017; 11:128-142. [PMID: 28113678 DOI: 10.1109/tbcas.2016.2579164] [Citation(s) in RCA: 33] [Impact Index Per Article: 4.7] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 06/06/2023]
Abstract
We present results from a new approach to learning and plasticity in neuromorphic hardware systems: to enable flexibility in implementable learning mechanisms while keeping high efficiency associated with neuromorphic implementations, we combine a general-purpose processor with full-custom analog elements. This processor is operating in parallel with a fully parallel neuromorphic system consisting of an array of synapses connected to analog, continuous time neuron circuits. Novel analog correlation sensor circuits process spike events for each synapse in parallel and in real-time. The processor uses this pre-processing to compute new weights possibly using additional information following its program. Therefore, to a certain extent, learning rules can be defined in software giving a large degree of flexibility. Synapses realize correlation detection geared towards Spike-Timing Dependent Plasticity (STDP) as central computational primitive in the analog domain. Operating at a speed-up factor of 1000 compared to biological time-scale, we measure time-constants from tens to hundreds of micro-seconds. We analyze variability across multiple chips and demonstrate learning using a multiplicative STDP rule. We conclude that the presented approach will enable flexible and efficient learning as a platform for neuroscientific research and technological applications.
Collapse
|
5
|
Petrovici MA, Vogginger B, Müller P, Breitwieser O, Lundqvist M, Muller L, Ehrlich M, Destexhe A, Lansner A, Schüffny R, Schemmel J, Meier K. Characterization and compensation of network-level anomalies in mixed-signal neuromorphic modeling platforms. PLoS One 2014; 9:e108590. [PMID: 25303102 PMCID: PMC4193761 DOI: 10.1371/journal.pone.0108590] [Citation(s) in RCA: 32] [Impact Index Per Article: 3.2] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 04/29/2014] [Accepted: 08/22/2014] [Indexed: 11/18/2022] Open
Abstract
Advancing the size and complexity of neural network models leads to an ever increasing demand for computational resources for their simulation. Neuromorphic devices offer a number of advantages over conventional computing architectures, such as high emulation speed or low power consumption, but this usually comes at the price of reduced configurability and precision. In this article, we investigate the consequences of several such factors that are common to neuromorphic devices, more specifically limited hardware resources, limited parameter configurability and parameter variations due to fixed-pattern noise and trial-to-trial variability. Our final aim is to provide an array of methods for coping with such inevitable distortion mechanisms. As a platform for testing our proposed strategies, we use an executable system specification (ESS) of the BrainScaleS neuromorphic system, which has been designed as a universal emulation back-end for neuroscientific modeling. We address the most essential limitations of this device in detail and study their effects on three prototypical benchmark network models within a well-defined, systematic workflow. For each network model, we start by defining quantifiable functionality measures by which we then assess the effects of typical hardware-specific distortion mechanisms, both in idealized software simulations and on the ESS. For those effects that cause unacceptable deviations from the original network dynamics, we suggest generic compensation mechanisms and demonstrate their effectiveness. Both the suggested workflow and the investigated compensation mechanisms are largely back-end independent and do not require additional hardware configurability beyond the one required to emulate the benchmark networks in the first place. We hereby provide a generic methodological environment for configurable neuromorphic devices that are targeted at emulating large-scale, functional neural networks.
Collapse
Affiliation(s)
- Mihai A. Petrovici
- Ruprecht-Karls-Universität Heidelberg, Kirchhoff Institute for Physics, Heidelberg, Germany
| | - Bernhard Vogginger
- Technische Universität Dresden, Institute of Circuits and Systems, Dresden, Germany
| | - Paul Müller
- Ruprecht-Karls-Universität Heidelberg, Kirchhoff Institute for Physics, Heidelberg, Germany
| | - Oliver Breitwieser
- Ruprecht-Karls-Universität Heidelberg, Kirchhoff Institute for Physics, Heidelberg, Germany
| | - Mikael Lundqvist
- Department of Computational Biology, School of Computer Science and Communication, Stockholm University and Royal Institute of Technology, Stockholm, Sweden
| | - Lyle Muller
- CNRS, Unité de Neuroscience, Information et Complexité, Gif sur Yvette, France
| | - Matthias Ehrlich
- Technische Universität Dresden, Institute of Circuits and Systems, Dresden, Germany
| | - Alain Destexhe
- CNRS, Unité de Neuroscience, Information et Complexité, Gif sur Yvette, France
| | - Anders Lansner
- Department of Computational Biology, School of Computer Science and Communication, Stockholm University and Royal Institute of Technology, Stockholm, Sweden
| | - René Schüffny
- Technische Universität Dresden, Institute of Circuits and Systems, Dresden, Germany
| | - Johannes Schemmel
- Ruprecht-Karls-Universität Heidelberg, Kirchhoff Institute for Physics, Heidelberg, Germany
| | - Karlheinz Meier
- Ruprecht-Karls-Universität Heidelberg, Kirchhoff Institute for Physics, Heidelberg, Germany
| |
Collapse
|
6
|
Pfeil T, Grübl A, Jeltsch S, Müller E, Müller P, Petrovici MA, Schmuker M, Brüderle D, Schemmel J, Meier K. Six networks on a universal neuromorphic computing substrate. Front Neurosci 2013; 7:11. [PMID: 23423583 PMCID: PMC3575075 DOI: 10.3389/fnins.2013.00011] [Citation(s) in RCA: 115] [Impact Index Per Article: 10.5] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/19/2012] [Accepted: 01/18/2013] [Indexed: 11/28/2022] Open
Abstract
In this study, we present a highly configurable neuromorphic computing substrate and use it for emulating several types of neural networks. At the heart of this system lies a mixed-signal chip, with analog implementations of neurons and synapses and digital transmission of action potentials. Major advantages of this emulation device, which has been explicitly designed as a universal neural network emulator, are its inherent parallelism and high acceleration factor compared to conventional computers. Its configurability allows the realization of almost arbitrary network topologies and the use of widely varied neuronal and synaptic parameters. Fixed-pattern noise inherent to analog circuitry is reduced by calibration routines. An integrated development environment allows neuroscientists to operate the device without any prior knowledge of neuromorphic circuit design. As a showcase for the capabilities of the system, we describe the successful emulation of six different neural networks which cover a broad spectrum of both structure and functionality.
Collapse
Affiliation(s)
- Thomas Pfeil
- Kirchhoff-Institute for Physics, Universität Heidelberg Heidelberg, Germany
| | | | | | | | | | | | | | | | | | | |
Collapse
|
7
|
Pfeil T, Potjans TC, Schrader S, Potjans W, Schemmel J, Diesmann M, Meier K. Is a 4-bit synaptic weight resolution enough? - constraints on enabling spike-timing dependent plasticity in neuromorphic hardware. Front Neurosci 2012; 6:90. [PMID: 22822388 PMCID: PMC3398398 DOI: 10.3389/fnins.2012.00090] [Citation(s) in RCA: 68] [Impact Index Per Article: 5.7] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/27/2012] [Accepted: 06/04/2012] [Indexed: 11/13/2022] Open
Abstract
Large-scale neuromorphic hardware systems typically bear the trade-off between detail level and required chip resources. Especially when implementing spike-timing dependent plasticity, reduction in resources leads to limitations as compared to floating point precision. By design, a natural modification that saves resources would be reducing synaptic weight resolution. In this study, we give an estimate for the impact of synaptic weight discretization on different levels, ranging from random walks of individual weights to computer simulations of spiking neural networks. The FACETS wafer-scale hardware system offers a 4-bit resolution of synaptic weights, which is shown to be sufficient within the scope of our network benchmark. Our findings indicate that increasing the resolution may not even be useful in light of further restrictions of customized mixed-signal synapses. In addition, variations due to production imperfections are investigated and shown to be uncritical in the context of the presented study. Our results represent a general framework for setting up and configuring hardware-constrained synapses. We suggest how weight discretization could be considered for other backends dedicated to large-scale simulations. Thus, our proposition of a good hardware verification practice may rise synergy effects between hardware developers and neuroscientists.
Collapse
Affiliation(s)
- Thomas Pfeil
- Kirchhoff Institute for Physics, Ruprecht-Karls-University Heidelberg Heidelberg, Germany
| | | | | | | | | | | | | |
Collapse
|
8
|
Democratic population decisions result in robust policy-gradient learning: a parametric study with GPU simulations. PLoS One 2011; 6:e18539. [PMID: 21572529 PMCID: PMC3087717 DOI: 10.1371/journal.pone.0018539] [Citation(s) in RCA: 11] [Impact Index Per Article: 0.8] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/01/2010] [Accepted: 03/03/2011] [Indexed: 11/28/2022] Open
Abstract
High performance computing on the Graphics Processing Unit (GPU) is an emerging
field driven by the promise of high computational power at a low cost. However,
GPU programming is a non-trivial task and moreover architectural limitations
raise the question of whether investing effort in this direction may be
worthwhile. In this work, we use GPU programming to simulate a two-layer network
of Integrate-and-Fire neurons with varying degrees of recurrent connectivity and
investigate its ability to learn a simplified navigation task using a
policy-gradient learning rule stemming from Reinforcement Learning. The purpose
of this paper is twofold. First, we want to support the use of GPUs in the field
of Computational Neuroscience. Second, using GPU computing power, we investigate
the conditions under which the said architecture and learning rule demonstrate
best performance. Our work indicates that networks featuring strong
Mexican-Hat-shaped recurrent connections in the top layer, where decision making
is governed by the formation of a stable activity bump in the neural population
(a “non-democratic” mechanism), achieve mediocre learning results at
best. In absence of recurrent connections, where all neurons “vote”
independently (“democratic”) for a decision via population vector
readout, the task is generally learned better and more robustly. Our study would
have been extremely difficult on a desktop computer without the use of GPU
programming. We present the routines developed for this purpose and show that a
speed improvement of 5x up to 42x is provided versus optimised Python code. The
higher speed is achieved when we exploit the parallelism of the GPU in the
search of learning parameters. This suggests that efficient GPU programming can
significantly reduce the time needed for simulating networks of spiking neurons,
particularly when multiple parameter configurations are investigated.
Collapse
|
9
|
Brüderle D, Petrovici MA, Vogginger B, Ehrlich M, Pfeil T, Millner S, Grübl A, Wendt K, Müller E, Schwartz MO, de Oliveira DH, Jeltsch S, Fieres J, Schilling M, Müller P, Breitwieser O, Petkov V, Muller L, Davison AP, Krishnamurthy P, Kremkow J, Lundqvist M, Muller E, Partzsch J, Scholze S, Zühl L, Mayr C, Destexhe A, Diesmann M, Potjans TC, Lansner A, Schüffny R, Schemmel J, Meier K. A comprehensive workflow for general-purpose neural modeling with highly configurable neuromorphic hardware systems. BIOLOGICAL CYBERNETICS 2011; 104:263-296. [PMID: 21618053 DOI: 10.1007/s00422-011-0435-9] [Citation(s) in RCA: 35] [Impact Index Per Article: 2.7] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Received: 11/10/2010] [Accepted: 04/19/2011] [Indexed: 05/30/2023]
Abstract
In this article, we present a methodological framework that meets novel requirements emerging from upcoming types of accelerated and highly configurable neuromorphic hardware systems. We describe in detail a device with 45 million programmable and dynamic synapses that is currently under development, and we sketch the conceptual challenges that arise from taking this platform into operation. More specifically, we aim at the establishment of this neuromorphic system as a flexible and neuroscientifically valuable modeling tool that can be used by non-hardware experts. We consider various functional aspects to be crucial for this purpose, and we introduce a consistent workflow with detailed descriptions of all involved modules that implement the suggested steps: The integration of the hardware interface into the simulator-independent model description language PyNN; a fully automated translation between the PyNN domain and appropriate hardware configurations; an executable specification of the future neuromorphic system that can be seamlessly integrated into this biology-to-hardware mapping process as a test bench for all software layers and possible hardware design modifications; an evaluation scheme that deploys models from a dedicated benchmark library, compares the results generated by virtual or prototype hardware devices with reference software simulations and analyzes the differences. The integration of these components into one hardware-software workflow provides an ecosystem for ongoing preparative studies that support the hardware design process and represents the basis for the maturity of the model-to-hardware mapping software. The functionality and flexibility of the latter is proven with a variety of experimental results.
Collapse
Affiliation(s)
- Daniel Brüderle
- Kirchhoff Institute for Physics, Ruprecht-Karls-Universität Heidelberg, Heidelberg, Germany.
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
Collapse
|