相关实验视频

Updated: Jun 16, 2026

A Simple Stimulatory Device for Evoking Point-like Tactile Stimuli: A Searchlight for LFP to Spike Transitions

A Simple Stimulatory Device for Evoking Point-like Tactile Stimuli: A Searchlight for LFP to Spike Transitions

Published on: March 25, 2014

高性能深尖神经网络,每个神经元有0.3个尖峰

Ana Stanojevic^1,2, Stanisław Woźniak³, Guillaume Bellec^2,4

¹IBM Research Europe - Zurich, Rüschlikon, Switzerland.

Nature communications

|August 9, 2024

概括

现在可以为节能AI训练深度尖端神经网络. 我们的方法克服了梯度问题,使尖端神经网络能够匹配人工神经网络的性能,在图像分类任务中具有最小的尖端.

科学领域:

计算神经科学是一种神经科学.
人工智能的人工智能
机器学习机器学习

背景情况:

生物大脑通过罕见的二进制尖峰来实现能源效率.
与人工神经网络 (ANN) 相比,训练生物启发的尖端神经网络 (SNN) 是一个挑战.
从ANN到SNN使用时间到第一个尖峰编码存在理论映射,但实际培训仍然很困难.

研究的目的:

分析时间到第一个峰值SNNs的学习动态.
在SNN培训中识别和解决消失/爆炸梯度问题.
展示深度SNNs的有效培训,以实现高性能图像分类.

主要方法:

关于SNN学习动态的理论分析.
模拟SNNs,重点关注神经元膜潜力在值.
使用强大的梯度下降算法进行训练和微调.
在图像数据集上进行测试,包括MNIST,时尚-MNIST,CIFAR10,CIFAR100和PLACES365.5.

主要成果:

在SNN中确定了一个特定的消失/爆炸梯度问题.
在值处恒定的膜电位倾斜率确保了与ANN (使用修正线性单位) 的训练等价性.
深度SNN在各种图像分类基准上实现了与ANN匹配的性能.
使用每神经元少于0.3个峰值,实现了高性能分类.

结论:

深度SNN可以从头开始有效地训练或微调以匹配ANN性能.
拟议的方法可以实现节能的人工智能实现.
优化的SNN显示出低延迟,耐噪声硬件应用的潜力.

更多相关视频

Closed-loop Neuro-robotic Experiments to Test Computational Properties of Neuronal Networks

Closed-loop Neuro-robotic Experiments to Test Computational Properties of Neuronal Networks

Published on: March 2, 2015

Using Neuron Spiking Activity to Trigger Closed-Loop Stimuli in Neurophysiological Experiments

Using Neuron Spiking Activity to Trigger Closed-Loop Stimuli in Neurophysiological Experiments

Published on: November 12, 2019

相关概念视频

Neuron Structure

Neuron Structure

Neuron Structure

Neuron Structure

Neurons are the main type of cell in the nervous system that generate and transmit electrochemical signals. They primarily communicate with each other using neurotransmitters at specific junctions called synapses. Neurons come in many shapes that often relate to their function, but most share three main structures: an axon and dendrites that extend out from a cell body.
Structure and Function of Neurons
The neuronal cell body—the soma— houses the nucleus and organelles vital to cellular...

Neural Circuits

Neural Circuits

Neural circuits and neuronal pools are two of the main structures found in the nervous system. Neural circuits are networks of neurons that work together to carry out a specific task or process. They consist of interconnected neurons and glial cells, which provide structural and metabolic support.
Neuronal pools are collections of nerve cells with similar functions and interact through chemical and electrical signals. These pools include both interneurons (the central neural circuit nodes that...