Satoshi Moriya
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Satoshi Moriya has a total of 46 co-authors. Here are the top co-authors by number of shared publications:
- Shigeo Sato
Research Institute of Electrical Communication, Tohoku University, Sendai 980-8577, Japan
15 shared publications - Hideaki Yamamoto
Tohoku University
12 shared publications - Ayumi Hirano-Iwata
Tohoku University
5 shared publications - Jordi Madrenas
Department of Electronic Engineering, Universitat Politècnica de Catalunya, Barcelona, Spain
4 shared publications - Yoshihiko Horio
ACRO, Teikyo University
4 shared publications Hisanao Akima
Research Institute of Electrical Communication, Tohoku University, Sendai, Miyagi 980-8577, Japan
3 shared publications
Latest Publications
- 15 September 2026
- Published by Springer Nature
- p. 284-294
Design of a Hardware-Efficient Digital Inference Circuit for SNN-Based Reservoir Computing Systems
- 15 September 2026
- Published by Springer Nature
- p. 284-294
Design of a Hardware-Efficient Digital Inference Circuit for SNN-Based Reservoir Computing Systems
- Open Access
- 24 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems I: Regular Papers
- Vol. 72 (10) , 5571-5582
Analog VLSI Implementation of Subthreshold Spiking Neural Networks and Its Application to Reservoir Computing
Neuromorphic computing achieves highly energy-efficient computations while adapting to environmental changes. Processing time series data by spiking neural networks can further reduce the power consumption of neuromorphic computing devices because most of the energy in the network is consumed only when the neuron generates and transmits a spike. In this study, we designed fully analog two-variable spiking neuron and spiking neural network circuits, taking advantage of the physical properties of transistors as analog devices. The energy consumption of the circuit for generating a spike was 22.7 fJ/spike when the MOS transistors were operating in the subthreshold region. The proposed circuits were implemented on an analog very large-scale integrated (VLSI) circuit chip in a 0.18 µm CMOS process. The circuits exhibit complex spike dynamics even under subthreshold operation according to chip measurements. We demonstrated that the spike sequence generated by the spiking neural network circuit was successfully applied to spoken digit recognition tasks via a reservoir computing framework with 14.4 fJ/SOP efficiency. These results provide important insights into edge AI applications of SNN-based neuromorphic hardware.
- 24 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems I: Regular Papers
- Vol. 72 (10) , 5571-5582
Open AccessAnalog VLSI Implementation of Subthreshold Spiking Neural Networks and Its Application to Reservoir Computing
Neuromorphic computing achieves highly energy-efficient computations while adapting to environmental changes. Processing time series data by spiking neural networks can further reduce the power consumption of neuromorphic computing devices because most of the energy in the network is consumed only when the neuron generates and transmits a spike. In this study, we designed fully analog two-variable spiking neuron and spiking neural network circuits, taking advantage of the physical properties of transistors as analog devices. The energy consumption of the circuit for generating a spike was 22.7 fJ/spike when the MOS transistors were operating in the subthreshold region. The proposed circuits were implemented on an analog very large-scale integrated (VLSI) circuit chip in a 0.18 µm CMOS process. The circuits exhibit complex spike dynamics even under subthreshold operation according to chip measurements. We demonstrated that the spike sequence generated by the spiking neural network circuit was successfully applied to spoken digit recognition tasks via a reservoir computing framework with 14.4 fJ/SOP efficiency. These results provide important insights into edge AI applications of SNN-based neuromorphic hardware.
- Open Access
- 28 November 2024
- Published by Elsevier in Neural Networks
- Vol. 184, 106967
Directional intermodular coupling enriches functional complexity in biological neuronal networks
- 28 November 2024
- Published by Elsevier in Neural Networks
- Vol. 184, 106967
Open AccessDirectional intermodular coupling enriches functional complexity in biological neuronal networks
- 8 January 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Applied Superconductivity
- Vol. 34 (3) , 1-4
Numerical Study on Physical Reservoir Computing With Josephson Junctions
In this study, we propose reservoir computing, a novel machine learning framework, utilizing the Josephson transmission line (JTL) as a promising hardware candidate to realize low-power and high-speed computation. A two-dimensional JTL circuit is designed as a reservoir in accordance with a previous study, and digit image recognition tasks are demonstrated with the circuit. The simulation results show that noisy digit images are successfully classified with an accuracy of $\text{80}\%$ at a rate of $\text{50}\, {\mathbf{Gpixels/s}}$ . The power consumption of this system is estimated to be $\text{12.8}\,\mu \text{W}$ , which is comparable to that of spin reservoirs and optical reservoirs. Thus, we confirm that the proposed system has great potential for application in machine learning and AI processing.
- 8 January 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Applied Superconductivity
- Vol. 34 (3) , 1-4
Numerical Study on Physical Reservoir Computing With Josephson Junctions
In this study, we propose reservoir computing, a novel machine learning framework, utilizing the Josephson transmission line (JTL) as a promising hardware candidate to realize low-power and high-speed computation. A two-dimensional JTL circuit is designed as a reservoir in accordance with a previous study, and digit image recognition tasks are demonstrated with the circuit. The simulation results show that noisy digit images are successfully classified with an accuracy of $\text{80}\%$ at a rate of $\text{50}\, {\mathbf{Gpixels/s}}$ . The power consumption of this system is estimated to be $\text{12.8}\,\mu \text{W}$ , which is comparable to that of spin reservoirs and optical reservoirs. Thus, we confirm that the proposed system has great potential for application in machine learning and AI processing.
- 22 September 2023
- Published by Springer Nature
- p. 423-434
Real-Time Adaptive Physical Sensor Processing with SNN Hardware
- 22 September 2023
- Published by Springer Nature
- p. 423-434
Real-Time Adaptive Physical Sensor Processing with SNN Hardware
- Open Access
Biological neurons act as generalization filters in reservoir computing
Reservoir computing is a machine learning paradigm that transforms the transient dynamics of high-dimensional nonlinear systems for processing time-series data. Although reservoir computing was initially proposed to model information processing in the mammalian cortex, it remains unclear how the non-random network architecture, such as the modular architecture, in the cortex integrates with the biophysics of living neurons to characterize the function of biological neuronal networks (BNNs). Here, we used optogenetics and fluorescent calcium imaging to record the multicellular responses of cultured BNNs and employed the reservoir computing framework to decode their computational capabilities. Micropatterned substrates were used to embed the modular architecture in the BNNs. We first show that modular BNNs can be used to classify static input patterns with a linear decoder and that the modularity of the BNNs positively correlates with the classification accuracy. We then used a timer task to verify that BNNs possess a short-term memory of ~1 s and finally show that this property can be exploited for spoken digit classification. Interestingly, BNN-based reservoirs allow transfer learning, wherein a network trained on one dataset can be used to classify separate datasets of the same category. Such classification was not possible when the input patterns were directly decoded by a linear decoder, suggesting that BNNs act as a generalization filter to improve reservoir computing performance. Our findings pave the way toward a mechanistic understanding of information processing within BNNs and, simultaneously, build future expectations toward the realization of physical reservoir computing systems based on BNNs.
Open AccessBiological neurons act as generalization filters in reservoir computing
Reservoir computing is a machine learning paradigm that transforms the transient dynamics of high-dimensional nonlinear systems for processing time-series data. Although reservoir computing was initially proposed to model information processing in the mammalian cortex, it remains unclear how the non-random network architecture, such as the modular architecture, in the cortex integrates with the biophysics of living neurons to characterize the function of biological neuronal networks (BNNs). Here, we used optogenetics and fluorescent calcium imaging to record the multicellular responses of cultured BNNs and employed the reservoir computing framework to decode their computational capabilities. Micropatterned substrates were used to embed the modular architecture in the BNNs. We first show that modular BNNs can be used to classify static input patterns with a linear decoder and that the modularity of the BNNs positively correlates with the classification accuracy. We then used a timer task to verify that BNNs possess a short-term memory of ~1 s and finally show that this property can be exploited for spoken digit classification. Interestingly, BNN-based reservoirs allow transfer learning, wherein a network trained on one dataset can be used to classify separate datasets of the same category. Such classification was not possible when the input patterns were directly decoded by a linear decoder, suggesting that BNNs act as a generalization filter to improve reservoir computing performance. Our findings pave the way toward a mechanistic understanding of information processing within BNNs and, simultaneously, build future expectations toward the realization of physical reservoir computing systems based on BNNs.
- 1 January 2022
- Published by Institute of Electronics, Information and Communications Engineers (IEICE) in Nonlinear Theory and Its Applications, IEICE
- Vol. 13 (2) , 427-433
An investigation of the relationship between numerical precision and performance of Q-learning for hardware implementation
Reinforcement learning is promising as a machine learning paradigm in edge computing. However, its high computational cost poses a challenge when implementing in devices with limited circuit resources and power consumption. In this study, we investigated the relationship between the bit-length of floating-point operations and the learning performance of the reinforcement learning algorithm. In the case of the FrozenLake maze problem, we found that the learning performance of 8-bit floating-point arithmetic decreased, while that of 16-bit floating-point arithmetic was comparable to that of 64-bit CPU arithmetic. Our results provide a practical guideline for designing a dedicated reinforcement learning hardware with minimum circuit resources and power consumption.
- 1 January 2022
- Published by Institute of Electronics, Information and Communications Engineers (IEICE) in Nonlinear Theory and Its Applications, IEICE
- Vol. 13 (2) , 427-433
An investigation of the relationship between numerical precision and performance of Q-learning for hardware implementation
Reinforcement learning is promising as a machine learning paradigm in edge computing. However, its high computational cost poses a challenge when implementing in devices with limited circuit resources and power consumption. In this study, we investigated the relationship between the bit-length of floating-point operations and the learning performance of the reinforcement learning algorithm. In the case of the FrozenLake maze problem, we found that the learning performance of 8-bit floating-point arithmetic decreased, while that of 16-bit floating-point arithmetic was comparable to that of 64-bit CPU arithmetic. Our results provide a practical guideline for designing a dedicated reinforcement learning hardware with minimum circuit resources and power consumption.
- 7 September 2021
- Published by Springer Nature
- p. 177-181
A Subthreshold Spiking Neuron Circuit Based on the Izhikevich Model
- 7 September 2021
- Published by Springer Nature
- p. 177-181
A Subthreshold Spiking Neuron Circuit Based on the Izhikevich Model
- 1 January 2021
- Published by Institute of Electronics, Information and Communications Engineers (IEICE) in Nonlinear Theory and Its Applications, IEICE
- Vol. 12 (4) , 685-694
Analog-circuit implementation of multiplicative spike-timing-dependent plasticity with linear decay
Neuromorphic engineering is a promising computing paradigm in next-generation information and communication technology. In particular, spiking neural networks are expected to reduce power consumption drastically owing to their event-driven operation. The spike-timing-dependent plasticity (STDP) rule, which learns from local spike-timing differences between spiking neurons, is a biologically plausible learning rule for spiking neural networks (SNNs). In this study, we designed and simulated an analog circuit that reproduces the multiplicative STDP rule, which is more flexible and adaptive to external signals. We also derived analytical expressions for the behavior of the proposed circuit. These results provide important insights for designing energy efficient neuromorphic devices for applications including edge computing.
- 1 January 2021
- Published by Institute of Electronics, Information and Communications Engineers (IEICE) in Nonlinear Theory and Its Applications, IEICE
- Vol. 12 (4) , 685-694
Analog-circuit implementation of multiplicative spike-timing-dependent plasticity with linear decay
Neuromorphic engineering is a promising computing paradigm in next-generation information and communication technology. In particular, spiking neural networks are expected to reduce power consumption drastically owing to their event-driven operation. The spike-timing-dependent plasticity (STDP) rule, which learns from local spike-timing differences between spiking neurons, is a biologically plausible learning rule for spiking neural networks (SNNs). In this study, we designed and simulated an analog circuit that reproduces the multiplicative STDP rule, which is more flexible and adaptive to external signals. We also derived analytical expressions for the behavior of the proposed circuit. These results provide important insights for designing energy efficient neuromorphic devices for applications including edge computing.
- Open Access
- 1 January 2020
- Published by Institute of Electronics, Information and Communications Engineers (IEICE) in Nonlinear Theory and Its Applications, IEICE
- Vol. 11 (4) , 590-600
Modular networks of spiking neurons for applications in time-series information processing
Spiking neural networks with complex spatiotemporal dynamics support efficient information processing of time-series signals. Here, we investigate the relationship between complexity of network dynamics and modular topology of networks using numerical simulations and discuss their effect on the classification performance of spoken-digit recognition tasks. The results show that modular networks generate spatially complex dynamics in which partially and globally synchronous bursts coexist. The classification rate of the modular reservoir network was approximately 75%, a value of which was comparable to that of a random network. This was caused by the randomly-connection structure between the input-reservoir and reservoir-readout layers, thus appropriate inference methods and asymmetry of connections should be introduced to take advantage of the complex dynamics in modular networks.
- 1 January 2020
- Published by Institute of Electronics, Information and Communications Engineers (IEICE) in Nonlinear Theory and Its Applications, IEICE
- Vol. 11 (4) , 590-600
Open AccessModular networks of spiking neurons for applications in time-series information processing
Spiking neural networks with complex spatiotemporal dynamics support efficient information processing of time-series signals. Here, we investigate the relationship between complexity of network dynamics and modular topology of networks using numerical simulations and discuss their effect on the classification performance of spoken-digit recognition tasks. The results show that modular networks generate spatially complex dynamics in which partially and globally synchronous bursts coexist. The classification rate of the modular reservoir network was approximately 75%, a value of which was comparable to that of a random network. This was caused by the randomly-connection structure between the input-reservoir and reservoir-readout layers, thus appropriate inference methods and asymmetry of connections should be introduced to take advantage of the complex dynamics in modular networks.