A Metaheuristic Approach for Parameter Fitting in Digital Spiking Silicon Neuron Model

Nanami, Takuya; Grassia, Filippo; Kohno, Takashi

doi:10.2991/jrnal.2018.5.1.8

Cited by 4 publications

(12 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The input current was given from the PC to the PQN unit on the FPGA via serial communication, and the value of the membrane potential v of the PQN unit was sent back to the PC. For comparison, we also prepared simulation results of DSSN models for each class; the DSSN model presented in Nanami et al ( 2017 , 2018 ) for the RS, FS, LTS, and IB classes, and Nanami et al ( 2016 ) for the EB and PB classes. The DSSN models and ionic-conductance models were simulated on a PC using the python software.…”

Section: Resultsmentioning

confidence: 99%

“…Table 1 shows resource consumption of the PQN, DSSN, and ionic-conductance models in the FPGA implementation. Previous studies (Nanami et al, 2016 , 2017 , 2018 ) of the DSSN model showed only the resources of the circuit of the DSSN engine, which computes the values of the state variables in the next step from the current values of the state variables. Therefore, we show the resources of both the PQN engine and the PQN unit for comparison.…”

Section: Resultsmentioning

confidence: 99%

“…In addition, buffers in the pipeline process consume more FF in the PQN model. Please note that the implementation of the DSSN model was not pipelined in the previous studies (Nanami et al, 2017(Nanami et al, , 2018 because the main goal of these studies was to build parameter fitting methods of the model and the implementation was done only to verify that it worked correctly. Therefore, the maximum clock frequency and the number of neurons that can be run in the real-time simulation were not measured.…”

Section: Discussionmentioning

confidence: 99%

“…The neuronal parameters are fitted so that the model reproduces the properties of each class as accurately as possible and consumes as few circuit resources as possible in the digital arithmetic circuit implementation. The parameters are determined by a semi-automatic optimization method developed previously, the detailed protocol of which is described in Nanami et al (2017Nanami et al ( , 2018. Since the synaptic current is not fitted in this study, the synaptic parameters α and β are the same values used in the previous study (Li et al, 2012).…”

Section: Equations Of the Pqn Modelmentioning

confidence: 99%

See 3 more Smart Citations

Piecewise quadratic neuron model: A tool for close-to-biology spiking neuronal network simulation on dedicated hardware

Nanami

Kohno

2023

Front. Neurosci.

Self Cite

View full text Add to dashboard Cite

Spiking neuron models simulate neuronal activities and allow us to analyze and reproduce the information processing of the nervous system. However, ionic-conductance models, which can faithfully reproduce neuronal activities, require a huge computational cost, while integral-firing models, which are computationally inexpensive, have some difficulties in reproducing neuronal activities. Here we propose a Piecewise Quadratic Neuron (PQN) model based on a qualitative modeling approach that aims to reproduce only the key dynamics behind neuronal activities. We demonstrate that PQN models can accurately reproduce the responses of ionic-conductance models of major neuronal classes to stimulus inputs of various magnitudes. In addition, the PQN model is designed to support the efficient implementation on digital arithmetic circuits for use as silicon neurons, and we confirm that the PQN model consumes much fewer circuit resources than the ionic-conductance models. This model intends to serve as a tool for building a large-scale closer-to-biology spiking neural network.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Resultsmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

Section: Equations Of the Pqn Modelmentioning

confidence: 99%

See 2 more Smart Citations

Piecewise quadratic neuron model: A tool for close-to-biology spiking neuronal network simulation on dedicated hardware

Nanami

Kohno

2023

Front. Neurosci.

Self Cite

View full text Add to dashboard Cite

show abstract

“…The system is a relatively small (∼2200 neurons) network having a known function, whose complete network topology, or connectome, is available. The electrophysiological activity of neurons was reproduced by using the piecewise quadratic neuron (PQN) model, which is a lightweight neuron model suitable for digital arithmetic circuit implementations [11][12][13][14][15][16].…”

Section: Introductionmentioning

confidence: 99%

A lightweight data-driven spiking neural network model ofDrosophilaolfactory nervous system with dedicated hardware support

Nanami,

Yamada,

Someya

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

Data-driven spiking neural network (SNN) models are vital for understanding the brain's information processing at the cellular and synaptic level. While extensive research has focused on developing data-driven SNN models for mammalian brains, their complexity poses challenges in achieving precision. Network topology often relies on statistical inference, and the functions of specific brain regions and supporting neuronal activities remain unclear. Additionally, these models demand significant computational resources. Here, we propose a lightweight data-driven SNN model that strikes a balance between simplicity and reproducibility. We target the Drosophila olfactory nervous system, extracting its network topology from connectome data. The model implemented on an entry-level field-programmable gate array successfully reproduced the functions and characteristic spiking activities of different neuron types. Our approach thus provides a foundation for constructing lightweight in silico models that are critical for investigating the brain's information processing mechanisms at the cellular and synaptic level through an analysis-by-construction approach and applicable to edge artificial intelligence (AI) systems.

show abstract

A lightweight data-driven spiking neuronal network model of Drosophila olfactory nervous system with dedicated hardware support

Nanami,

Yamada,

Someya

et al. 2024

Front. Neurosci.

Self Cite

View full text Add to dashboard Cite

Data-driven spiking neuronal network (SNN) models enable in-silico analysis of the nervous system at the cellular and synaptic level. Therefore, they are a key tool for elucidating the information processing principles of the brain. While extensive research has focused on developing data-driven SNN models for mammalian brains, their complexity poses challenges in achieving precision. Network topology often relies on statistical inference, and the functions of specific brain regions and supporting neuronal activities remain unclear. Additionally, these models demand huge computing facilities and their simulation speed is considerably slower than real-time. Here, we propose a lightweight data-driven SNN model that strikes a balance between simplicity and reproducibility. The model is built using a qualitative modeling approach that can reproduce key dynamics of neuronal activity. We target the Drosophila olfactory nervous system, extracting its network topology from connectome data. The model was successfully implemented on a small entry-level field-programmable gate array and simulated the activity of a network in real-time. In addition, the model reproduced olfactory associative learning, the primary function of the olfactory system, and characteristic spiking activities of different neuron types. In sum, this paper propose a method for building data-driven SNN models from biological data. Our approach reproduces the function and neuronal activities of the nervous system and is lightweight, acceleratable with dedicated hardware, making it scalable to large-scale networks. Therefore, our approach is expected to play an important role in elucidating the brain's information processing at the cellular and synaptic level through an analysis-by-construction approach. In addition, it may be applicable to edge artificial intelligence systems in the future.

show abstract

A Metaheuristic Approach for Parameter Fitting in Digital Spiking Silicon Neuron Model

Cited by 4 publications

References 10 publications

Piecewise quadratic neuron model: A tool for close-to-biology spiking neuronal network simulation on dedicated hardware

Piecewise quadratic neuron model: A tool for close-to-biology spiking neuronal network simulation on dedicated hardware

A lightweight data-driven spiking neural network model ofDrosophilaolfactory nervous system with dedicated hardware support

A lightweight data-driven spiking neuronal network model of Drosophila olfactory nervous system with dedicated hardware support

Contact Info

Product

Resources

About