Memristor-Based Edge Computing of ShuffleNetV2 for Image Classification

Ran, Huanhuan; Wen, Shiping; Wang, Shiqin; Cao, Yuting; Zhou, Pan; Huang, Tingwen

doi:10.1109/tcad.2020.3022970

Cited by 18 publications

(9 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Traditional digital signal processing methods usually use analog signal sensors and digital processing units which are physically segmented from each other, leading to high latency and energy consumption for data transferring. IMCbased image processing has shown significant advantages not only in the parallel processing of MVM operations but also in taking advantage of the analog computing nature by reducing the distance between the signal sensing and processing in abundant techniques [121][122][123][124]. Furthermore, many new types of memristors, as well as other emerging nanodevices, have been developed as visual [125,126], auditory [127], and olfactory [128] sensors, etc., generating excitation from the applied signals and processing them in situ, and greatly improving the compactness, processing speed, and energy efficiency.…”

Section: Digital Image Processingmentioning

confidence: 99%

Toward memristive in-memory computing: principles and applications

Bao

Zhou

et al. 2022

Front. Optoelectron.

View full text Add to dashboard Cite

With the rapid growth of computer science and big data, the traditional von Neumann architecture suffers the aggravating data communication costs due to the separated structure of the processing units and memories. Memristive in-memory computing paradigm is considered as a prominent candidate to address these issues, and plentiful applications have been demonstrated and verified. These applications can be broadly categorized into two major types: soft computing that can tolerant uncertain and imprecise results, and hard computing that emphasizes explicit and precise numerical results for each task, leading to different requirements on the computational accuracies and the corresponding hardware solutions. In this review, we conduct a thorough survey of the recent advances of memristive in-memory computing applications, both on the soft computing type that focuses on artificial neural networks and other machine learning algorithms, and the hard computing type that includes scientific computing and digital image processing. At the end of the review, we discuss the remaining challenges and future opportunities of memristive in-memory computing in the incoming Artificial Intelligence of Things era. Graphical Abstract

show abstract

Section: Digital Image Processingmentioning

confidence: 99%

Toward memristive in-memory computing: principles and applications

Bao

Zhou

et al. 2022

Front. Optoelectron.

View full text Add to dashboard Cite

show abstract

“…Our memristor-based MobileNetV3 neural network achieves a high classification accuracy of 90.36% on the CIFAR-10 dataset, comparable to traditional implementations using the PyTorch framework and exhibiting a significant improvement over previous works in this area. We refer to methods previously reported in the papers to analyze the latency and energy consumption of memristor-based neural networks (Wen et al, 2019;Ran et al, 2020;Yang et al, 2022;Zhang et al, 2023), to determine the latency of our proposed computing paradigm. Appendix F provides an analysis of the resources required for the memristor-based MobileNetV3 in the CIFAR-10 image classification task.…”

Section: Accuracymentioning

confidence: 99%

“…Appendix F provides an analysis of the resources required for the memristor-based MobileNetV3 in the CIFAR-10 image classification task. In neural circuits based on memristors, the response time can be as quick as 100 ps (Ran et al, 2020). However, the overall speed of the memristor-based circuit is constrained by the slew rate of the op-amps employed for converting…”

Section: Accuracymentioning

confidence: 99%

Memristor-Based In-Memory Computing Architecture for Scientific Computing

Li²,

Ling³

et al. 2022

Memristor Computing Systems

View full text Add to dashboard Cite

The advancement in the field of machine learning is inextricably linked with the concurrent progress in domain-specific hardware accelerators such as GPUs and TPUs. However, the rapidly growing computational demands necessitated by larger models and increased data have become a primary bottleneck in further advancing machine learning, especially in mobile and edge devices. Currently, the neuromorphic computing paradigm based on memristors presents a promising solution. In this study, we introduce a memristorbased MobileNetV3 neural network computing paradigm and provide an end-to-end framework for validation. The results demonstrate that this computing paradigm achieves over 90% accuracy on the CIFAR-10 dataset while saving inference time and reducing energy consumption. With the successful development and verification of MobileNetV3, the potential for realizing more memristor-based neural networks using this computing paradigm and open-source framework has significantly increased. This progress sets a groundbreaking pathway for future deployment initiatives.

show abstract

“…A memristor crossbar in figure 3(c) can be one promising candidate for implementing energy-efficient and low-precision AI hardware including edge-intelligence (Keshavarzi and van den Hoek 2019, Krestinskaya et al 2019, Zhou et al 2019, Deng et al 2020, Keshavarzi et al 2020, Ran et al 2020, Xue et al 2020, Qin et al 2020a, Singh et al 2021. In-memory computing with the memristor crossbar in figure 3(c) can be used to overcome the von Neumann machine's memory access bottleneck mentioned earlier.…”

Section: Introductionmentioning

confidence: 99%

Quantization, training, parasitic resistance correction, and programming techniques of memristor-crossbar neural networks for edge intelligence

Nguyen

et al. 2022

Neuromorph. Comput. Eng.

View full text Add to dashboard Cite

In Internet-of-Things (IoT) era, edge intelligence is critical for overcoming the communication and computing energy crisis, which is unavoidable if cloud computing is used exclusively. Memristor crossbars with in-memory computing may be suitable for realizing edge intelligence hardware. They can perform both memory and computing functions, allowing for the development of low-power computing architectures that go beyond the Von Neumann computer. For implementing edge-intelligence hardware with memristor crossbars, in this paper, we review various techniques such as quantization, training, parasitic resistance correction, and low-power crossbar programming, and so on. In particular, memristor crossbars can be considered to realize quantized neural networks with binary and ternary synapses. For preventing memristor defects from degrading edge intelligence performance, chip-in-the-loop training can be useful when training memristor crossbars. Another undesirable effect in memristor crossbars is parasitic resistances such as source, line, and neuron resistance, which worsens as crossbar size increases. Various circuit and software techniques can compensate for parasitic resistances like source, line, and neuron resistance. Finally, we discuss an energy-efficient programming method for updating synaptic weights in memristor crossbars, which is needed for learning the edge devices.

show abstract

Memristor-Based Edge Computing of ShuffleNetV2 for Image Classification

Cited by 18 publications

References 35 publications

Toward memristive in-memory computing: principles and applications

Toward memristive in-memory computing: principles and applications

Memristor-Based In-Memory Computing Architecture for Scientific Computing

Quantization, training, parasitic resistance correction, and programming techniques of memristor-crossbar neural networks for edge intelligence

Contact Info

Product

Resources

About