An Ultra-Low Power Threshold Voltage Variable Artificial Retina Neuron

Wang, Qiguang; Pan, Guangchen; Jiang, Yanfeng

doi:10.3390/electronics11030365

Cited by 3 publications

(1 citation statement)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Their primary advantages include low energy consumption as demonstrated for ZnObased memristor [7], high parallelism as demonstrated for image skeletonizing [8], signal processing efficiency as shown in [9] for a device with a very low switching current level and self-rectifying characteristics that can be utilized for reservoir computing and real-time adaptability as was discussed in [10]. Efficient solutions for a set of tasks, such as sensory information processing as demonstrated for novel artificial retinal neuron with ultra-low power in [11], pattern analysis as for Fashion MNIST dataset [12], speech recognition, e.g., for the Texas Instruments digit sequences dataset in [13], the Heidelberg Digits dataset and the Speech Commands dataset [14], and other perception and environment interactionrelated activities as for reinforcement learning in [15] can be achieved by neuromorphic processors. In some cases, this requires the application of methods based on the utilization of a limited number of trainable synaptic weights, which necessitates the development of spiking neural network topologies with limited plasticity synapses [16].…”

Section: Introductionmentioning

confidence: 99%

Extraction of Significant Features by Fixed-Weight Layer of Processing Elements for the Development of an Efficient Spiking Neural Network Classifier

Sboev,

Rybka,

Kunitsyn

et al. 2023

BDCC

View full text Add to dashboard Cite

In this paper, we demonstrate that fixed-weight layers generated from random distribution or logistic functions can effectively extract significant features from input data, resulting in high accuracy on a variety of tasks, including Fisher’s Iris, Wisconsin Breast Cancer, and MNIST datasets. We have observed that logistic functions yield high accuracy with less dispersion in results. We have also assessed the precision of our approach under conditions of minimizing the number of spikes generated in the network. It is practically useful for reducing energy consumption in spiking neural networks. Our findings reveal that the proposed method demonstrates the highest accuracy on Fisher’s iris and MNIST datasets with decoding using logistic regression. Furthermore, they surpass the accuracy of the conventional (non-spiking) approach using only logistic regression in the case of Wisconsin Breast Cancer. We have also investigated the impact of non-stochastic spike generation on accuracy.

show abstract

Section: Introductionmentioning

confidence: 99%

Extraction of Significant Features by Fixed-Weight Layer of Processing Elements for the Development of an Efficient Spiking Neural Network Classifier

Sboev,

Rybka,

Kunitsyn

et al. 2023

BDCC

View full text Add to dashboard Cite

show abstract

Ultra-Efficient Low-Power Retinal Nano Electronic Circuit for Edge Enhancement and Detection Using 7 nm FinFET Technology

Turiqul Islam,

Al-Shidaifat,

Jooq

et al. 2024

Journal of Nanoelectronics and Optoelectronics

View full text Add to dashboard Cite

This study proposed a 7 nm FinFET based analog one pixel circuit block inspired by lateral inhibition phenomenon to perform edge enhancing and edge detection of optoelectronic image. This plays a crucial role in retinomorphic applications like artificial human retinal functions. Proposed Edge enhancement and edge detection circuits are constructed using two distinct 750×750-pixel silicon networks. First the single pixel circuit cell is reconstructed with the lateral inhibition phenomenon, then the circuit using GPDK (Generic Process Design Kit) in 180 nm, 90 nm, and 45 nm CMOS technology is designed. We used 3×3 convolution process for image masking in digital and analog image signal processing which gives more accuracy in term of object recognition. The power consumption in each case is obtained to be approximately 19.71 μW, 4.18 μW and 1.62 μW for edge enhancing and 23.76 μW, 7.99 μW and 3.41 μW for edge detection which is much larger than the power consumed by the same circuit is implemented with 7 nm FinFET (Fin Field Effect Transistor) technology, 21.91 pW and 24.85 pW. In addition, the size reduction of the circuit reduced by 84% compared with 45 nm CMOS, increases the accuracy of the circuit by 30%. Results confirm that FinFET based single pixel circuit consumes less power, reduces size, and gives higher accuracy. The output from all the circuits has been matched with the biological response.

show abstract

Extraction of Significant Features by Fixed-Weight Layer of Processing Elements for the Development of an Effective Spiking Neural Network Classifier

Sboev,

Rybka,

Kunitsyn

et al. 2023

Preprint

View full text Add to dashboard Cite

In this paper, we demonstrate that fixed-weight layers, generated from random distribution or logistic functions, can effectively extract significant features from input data, resulting in high accuracy on a variety of tasks, including Fisher’s iris, Wisconsin Breast Cancer, and MNIST datasets. We have observed that logistic functions yield high accuracy with less dispersion in results. We have also assessed the precision of our approach under conditions of minimizing spikes number generated in the network, it is a practically useful for reducing energy consumption in spiking neural networks. Our findings reveal that the proposed method demonstrates a highest accuracy on Fisher’s iris and MNIST datasets with logistic regression. Furthermore, they surpass the accuracy of the conventional (non-spiking) approach using logistic regression in the case of Wisconsin Breast Cancer. We have also investigated the impact of non-stochastic spike generation on accuracy.

show abstract

An Ultra-Low Power Threshold Voltage Variable Artificial Retina Neuron

Cited by 3 publications

References 23 publications

Extraction of Significant Features by Fixed-Weight Layer of Processing Elements for the Development of an Efficient Spiking Neural Network Classifier

Extraction of Significant Features by Fixed-Weight Layer of Processing Elements for the Development of an Efficient Spiking Neural Network Classifier

Ultra-Efficient Low-Power Retinal Nano Electronic Circuit for Edge Enhancement and Detection Using 7 nm FinFET Technology

Extraction of Significant Features by Fixed-Weight Layer of Processing Elements for the Development of an Effective Spiking Neural Network Classifier

Contact Info

Product

Resources

About