Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Related Concept Videos

Lossy Lines and Overvoltages

Lossy Lines and Overvoltages

Transmission-line series resistance and shunt conductance cause three primary effects: attenuation, distortion, and power losses.
Attenuation
When constant series resistance and shunt conductance are present, voltage and current equations are modified. The propagation constant indicates that voltage and current waves consist of both forward and backward traveling components. These waves attenuate as they propagate, with the attenuation factor related to the resistance and conductance. In a...

Neural Regulation

Neural Regulation

Digestion begins with a cephalic phase that prepares the digestive system to receive food. When our brain processes visual or olfactory information about food, it triggers impulses in the cranial nerves innervating the salivary glands and stomach to prepare for food.

Linear Approximation in Frequency Domain

Linear Approximation in Frequency Domain

Linear systems are characterized by two main properties: superposition and homogeneity. Superposition allows the response to multiple inputs to be the sum of the responses to each individual input. Homogeneity ensures that scaling an input by a scalar results in the response being scaled by the same scalar.
In contrast, nonlinear systems do not inherently possess these properties. However, for small deviations around an operating point, a nonlinear system can often be approximated as linear....

The Role of Ion Channels in Neuronal Computation

The Role of Ion Channels in Neuronal Computation

A postsynaptic neuron usually receives numerous impulses from several other presynaptic neurons. The axon hillock of the postsynaptic neuron integrates all these signals and determines the likelihood of firing an action potential.
Sometimes a single EPSP is strong enough to induce an action potential in the postsynaptic neuron. However, multiple presynaptic inputs must often create EPSPs around the same time for the postsynaptic neuron to be sufficiently depolarized to fire an action potential....

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

Self-frequency-conversion nanowire lasers.

Light, science & applications·2022

Same author

The Effect of CD226 on the Balance between Inflammatory Monocytes and Small Peritoneal Macrophages in Mouse Ulcerative Colitis.

Immunological investigations·2022

Same author

MicroRNA-Messenger RNA Regulatory Network Mediates Disrupted TH17 Cell Differentiation in Depression.

Frontiers in psychiatry·2022

Same author

Tribocatalytic degradation of dyes by tungsten bronze ferroelectric Ba<sub>2.5</sub>Sr<sub>2.5</sub>Nb<sub>8</sub>Ta<sub>2</sub>O<sub>30</sub> submicron particles.

RSC advances·2022

Same author

NIR-I-Responsive Single-Band Upconversion Emission through Energy Migration in Core-Shell-Shell Nanostructures.

Angewandte Chemie (International ed. in English)·2022

Same author

Drought exerts a greater influence than growth temperature on the temperature response of leaf day respiration in wheat (Triticum aestivum).

Plant, cell & environment·2022

Search research articles

Related Experiment Video

Updated: Sep 21, 2025

Scalable Quantum Integrated Circuits on Superconducting Two-Dimensional Electron Gas Platform

Scalable Quantum Integrated Circuits on Superconducting Two-Dimensional Electron Gas Platform

Published on: August 2, 2019

Conductance-Aware Quantization Based on Minimum Error Substitution for Non-Linear-Conductance-State Tolerance in

Chenglong Huang¹, Nuo Xu^2,3, Wenqing Wang¹

¹Institute for Quantum Information & State Key Laboratory of High Performance Computing, National University of Defense Technology, Changsha 410073, China.

|May 28, 2022

Summary

This study introduces a conductance-aware quantization method to improve the accuracy of deep neural networks (DNNs) using resistive random-access memory (ReRAM) crossbar arrays. The new method effectively addresses non-linear ReRAM device behavior, enhancing in-memory computing performance.

Keywords:

ReRAM conductance-aware quantization non-linear conductance levels

More Related Videos

Identification and Classification of Position-specific GABAA Receptor Subunit Missense Variants for Their Role In Hippocampal Pyramidal Neurons

Identification and Classification of Position-specific GABA_A Receptor Subunit Missense Variants for Their Role In Hippocampal Pyramidal Neurons

Published on: June 6, 2025

Implementing Dynamic Clamp with Synaptic and Artificial Conductances in Mouse Retinal Ganglion Cells

Implementing Dynamic Clamp with Synaptic and Artificial Conductances in Mouse Retinal Ganglion Cells

Published on: May 16, 2013

Related Experiment Videos

Last Updated: Sep 21, 2025

Scalable Quantum Integrated Circuits on Superconducting Two-Dimensional Electron Gas Platform

Scalable Quantum Integrated Circuits on Superconducting Two-Dimensional Electron Gas Platform

Published on: August 2, 2019

Identification and Classification of Position-specific GABAA Receptor Subunit Missense Variants for Their Role In Hippocampal Pyramidal Neurons

Identification and Classification of Position-specific GABA_A Receptor Subunit Missense Variants for Their Role In Hippocampal Pyramidal Neurons

Published on: June 6, 2025

Implementing Dynamic Clamp with Synaptic and Artificial Conductances in Mouse Retinal Ganglion Cells

Implementing Dynamic Clamp with Synaptic and Artificial Conductances in Mouse Retinal Ganglion Cells

Published on: May 16, 2013

Area of Science:

Computer Science
Electrical Engineering
Materials Science

Background:

Resistive random-access memory (ReRAM) offers a path towards in-memory computing, aiming to overcome the limitations of traditional von Neumann architectures.
ReRAM crossbar arrays (RCAs) are crucial for accelerating deep neural network (DNN) computations, particularly multiplication-and-accumulation (MAC) operations.
Non-linear conductance distribution in ReRAM devices causes significant deviations when mapping quantized weights, leading to reduced inference accuracy in RCA-based DNNs.

Purpose of the Study:

To develop a quantization method that mitigates accuracy degradation caused by ReRAM's non-linear conductance characteristics.
To propose a solution that is adaptable to various ReRAM devices and robust against device variations.

Main Methods:

A minimum error substitution technique based on conductance-aware quantization was developed.
The proposed method quantizes weights considering the non-linear conductance distribution of ReRAM devices.
The method was evaluated on established DNN models like LeNet5, AlexNet, and VGG16.

Main Results:

The conductance-aware quantization method significantly reduces the deviation between ideal weights and actual ReRAM conductance values.
The proposed method demonstrates resilience to different non-linear conductance profiles and device variations.
Simulations confirmed substantial accuracy recovery for RCA-based DNNs compared to traditional linear quantization methods.

Conclusions:

The proposed minimum error substitution and conductance-aware quantization method effectively resolves the accuracy degradation issue in ReRAM-based DNNs.
This approach enhances the practical viability of ReRAM for efficient in-memory computing applications.
The method offers a robust solution for deploying DNNs on ReRAM hardware, improving inference accuracy and reliability.