Enabling a New Methodology of Neural Coding: Multiplexing Temporal Encoding in Neuromorphic Computing

Zheng, Honghao; Bai, Kang Jun; Yi, Yang

doi:10.1109/tvlsi.2023.3234514

Cited by 7 publications

(3 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Thus the relevance of subsequent spikes is insignificant. Researchers have suggested that using lesser spikes helps in power efficiency of a neuromorphic hardware [22], [23], [24], [25]. Therefore, the present approach should be power efficient.…”

Section: Impact and Insightsmentioning

confidence: 98%

See 1 more Smart Citation

Probabilistic Boolean Operation in Networks of Spiking Neurons for Bayesian Inference of Hidden Markov Models

Chakraborty¹,

Chakrabarti²

2023

Preprint

View full text Add to dashboard Cite

<p>Spiking neural networks (SNN) have gained popularity among the researchers for low power consumption. An interesting problem on Bayesian inference of hidden Markov models (HMM) on SNN paradigm is timely. In this work we propose a novel approach to address this issue. A probabilistic temporal encoding scheme is adopted and it has been shown that a spiking neuron behaves as a probabilistic Boolean operator. Using this property the posterior of a hidden state is mapped to probability of firing a logic $1$ by a spiking neuron. It has been demonstrated that both binary state HMM as well as multi-state HMMs are implementable using the present approach. The proposed method has been demonstrated to be effective for reasonably accurate estimation of hidden states. The present method seems to be reliable, power efficient and easy to follow.</p>

show abstract

Section: Impact and Insightsmentioning

confidence: 98%

“…All of these papers suggest adopting Poisson spike trains and rate coding. However, rate coding requires multiple spikes and therefore is prone to high energy dissipation [22], [23], [24], [25].…”

Section: Introductionmentioning

confidence: 99%

Probabilistic Boolean Operation in Networks of Spiking Neurons for Bayesian Inference of Hidden Markov Models

Chakraborty¹,

Chakrabarti²

2023

Preprint

View full text Add to dashboard Cite

show abstract

“…More synapse and LIF neuron circuits using different design scenarios can be found in [50]- [52]. In this work, by proposing a novel design method, we propose a compact and energy-efficient STDP and LIF neuron circuit using 7nm FinFET technology.…”

Section: B Previous Related Workmentioning

confidence: 99%

High-Performance and Energy-Efficient Leaky Integrate-and-Fire Neuron and Spike Timing-Dependent Plasticity Circuits in 7nm FinFET Technology

Jooq,

Azghadi,

Behbahani

et al. 2023

IEEE Access

View full text Add to dashboard Cite

In designing neuromorphic circuits and systems, developing compact and energy-efficient neuron and synapse circuits is essential for high-performance on-chip neural architectures. Toward that end, this work utilizes the advanced low-power and compact 7nm FinFET technology to design leaky integrateand-fire (LIF) neuron and spike-timing-dependent plasticity (STDP) circuits. In the proposed STDP circuit, only six FinFETs and three small capacitors (two 10fF and 20fF) have been utilized to realize STDP learning. Moreover, 12 transistors and two capacitors (20fF) have been employed for designing the LIF neuron circuit. The evaluation results demonstrate that besides 60% area saving, the proposed STDP circuit achieves 68% improvement in total average power consumption and 43% lower energy dissipation compared to previous works. The proposed LIF neuron circuit demonstrates 34% area saving, 46% power, and 40% energy saving compared to its counterparts. The neuron can also tune the firing frequency within 5MHz-330MHz using an external control voltage. These results emphasize the potential of the proposed neuron and STDP learning circuits for compact and energy-efficient neuromorphic computing systems.

show abstract

Enhancing SNN Training Performance: A Mixed-Signal Triplet Reconfigurable STDP Circuit with Multiplexing Encoding

Zheng

2023

2023 IEEE International Symposium on Circuits and Systems (ISCAS)

View full text Add to dashboard Cite

Enabling a New Methodology of Neural Coding: Multiplexing Temporal Encoding in Neuromorphic Computing

Cited by 7 publications

References 37 publications

Probabilistic Boolean Operation in Networks of Spiking Neurons for Bayesian Inference of Hidden Markov Models

Probabilistic Boolean Operation in Networks of Spiking Neurons for Bayesian Inference of Hidden Markov Models

High-Performance and Energy-Efficient Leaky Integrate-and-Fire Neuron and Spike Timing-Dependent Plasticity Circuits in 7nm FinFET Technology

Enhancing SNN Training Performance: A Mixed-Signal Triplet Reconfigurable STDP Circuit with Multiplexing Encoding

Contact Info

Product

Resources

About