人工神经网络ANN.ppt
合集下载
相关主题
- 1、下载文档前请自行甄别文档内容的完整性,平台不提供额外的编辑、内容补充、找答案等附加服务。
- 2、"仅部分预览"的文档,不可在线预览部分如存在完整性等问题,可反馈申请退款(可完整预览的文档不适用该条件!)。
- 3、如文档侵犯您的权益,请联系客服反馈,我们会尽快为您处理(人工客服工作时间:9:00-18:30)。
4wk.baidu.com
Biological Neural Networks
10/10/2020
• 10,000 synapses per neuron
• Computational power = connectivity
• Plasticity
– new connections (?)
– strength of connections modified
• 10 billion neurons in human brain • Summation of input stimuli
– Spatial (signals) – Temporal (pulses)
• Threshold over composed inputs • Constant firing strength
Artificial Neural Networks - I
5
Neural Dynamics
40 mV 20
membrane rest activation
0
Action potential
-20
-40 -60 -80 -100 -120
0
Refractory time
Action potential ≈ 100mV Activation threshold ≈ 20-30mV Rest potential ≈ -65mV Spike time ≈ 1-2ms Refractory time ≈ 10-20ms
wi2
wi3
wi4
j wij xj (t)
yi f (ui (t))
wi5
Neuron i
urest = resting potential xj(t) = output of neuron j at time t wij = connection strength between neuron i and neuron j u(t) = total stimulus at time t
• Unsupervised ANNs – Examples – Applications – Further topics
II
III
10/10/2020
Artificial Neural Networks - I
2
Contents - I
• Introduction to ANNs
– Processing elements (neurons) – Architecture
• 106 billion synapses in human brain
• Chemical transmission and modulation of signals
• Inhibitory synapses • Excitatory synapses
10/10/2020
Artificial Neural Networks - I
Artificial Neural Networks 人工神经网络
Introduction
Table of Contents
• Introduction to ANNs – Taxonomy – Features – Learning – Applications
I
• Supervised ANNs – Examples – Applications – Further topics
ms
10 20 30 40 50 60 70 80 90 100
10/10/2020
Artificial Neural Networks - I
6
神经网络的复杂性
• 神经网路的复杂多样,不仅在于神经元和突触 的数量大、组合方式复杂和联系广泛,还在于 突触传递的机制复杂。现在已经发现和阐明的 突触传递机制有:突触后兴奋,突触后抑制, 突触前抑制,突触前兴奋,以及“远程”抑制 等等。在突触传递机制中,释放神经递质是实 现突触传递机能的中心环节,而不同的神经递 质有着不同的作用性质和特点
• Functional Taxonomy of ANNs • Structural Taxonomy of ANNs • Features • Learning Paradigms • Applications
10/10/2020
Artificial Neural Networks - I
3
The Biological Neuron
10/10/2020
Artificial Neural Networks - I
7
神经网络的研究
• 神经系统活动,不论是感觉、运动,还是脑的 高级功能(如学习、记忆、情绪等)都有整体 上的表现,面对这种表现的神经基础和机理的 分析不可避免地会涉及各种层次。这些不同层 次的研究互相启示,互相推动。在低层次(细 胞、分子水平)上的工作为较高层次的观察提 供分析的基础,而较高层次的观察又有助于引 导低层次工作的方向和体现其功能意义。既有 物理的、化学的、生理的、心理的分门别类研 究,又有综合研究。
10/10/2020
Artificial Neural Networks - I
8
The Artificial Neuron
Stimulus
ui t wij x j t
j
Response
yi t f urest ui t
x1(t) x2(t) x3(t) x4(t) x5(t)
wi1
10/10/2020
Artificial Neural Networks - I
yi(t)
9
Artificial Neural Models
• McCulloch Pitts-type Neurons (static)
– Digital neurons: activation state interpretation (snapshot of the system each time a unit fires)
– Analog neurons: firing rate interpretation (activation of units equal to firing rate)
– Activation of neurons encodes information
• Spiking Neurons (dynamic)