版权说明:本文档由用户提供并上传,收益归属内容提供方,若内容存在侵权,请进行举报或认领
文档简介
1、Artificial Neural Networks人工神经网络,Introduction,19/12/2020,Artificial Neural Networks - I,2,Table of Contents,Introduction to ANNs Taxonomy Features Learning Applications,I,19/12/2020,Artificial Neural Networks - I,3,Contents - I,Introduction to ANNs Processing elements (neurons) Architecture Function
2、al Taxonomy of ANNs Structural Taxonomy of ANNs Features Learning Paradigms Applications,19/12/2020,Artificial Neural Networks - I,4,The Biological Neuron,10 billion neurons in human brain Summation of input stimuli Spatial (signals) Temporal (pulses) Threshold over composed inputs Constant firing s
3、trength,billion synapses in human brain Chemical transmission and modulation of signals Inhibitory synapses Excitatory synapses,19/12/2020,Artificial Neural Networks - I,5,Biological Neural Networks,10,000 synapses per neuron Computational power = connectivity Plasticity new connections (?) strength
4、 of connections modified,19/12/2020,Artificial Neural Networks - I,6,Neural Dynamics,Refractory time,Action potential,Action potential 100mV Activation threshold 20-30mV Rest potential -65mV Spike time 1-2ms Refractory time 10-20ms,19/12/2020,Artificial Neural Networks - I,7,神经网络的复杂性,神经网路的复杂多样,不仅在于神
5、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,19/12/2020,Artificial Neural Networks - I,8,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
6、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。,19/12/2020,Artificial Neural Networks - I,9,The Artificial Neuron,Stimulus,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
7、time t,yi(t),x1(t),x2(t),x5(t),x3(t),x4(t),wi1,wi3,wi2,wi4,wi5,Neuron i,Response,19/12/2020,Artificial Neural Networks - I,10,Artificial Neural Models,McCulloch Pitts-type Neurons (static) Digital neurons: activation state interpretation (snapshot of the system each time a unit fires) Analog neurons
8、: firing rate interpretation (activation of units equal to firing rate) Activation of neurons encodes information Spiking Neurons (dynamic) Firing pattern interpretation (spike trains of units) Timing of spike trains encodes information (time to first spike, phase of signal, correlation and synchron
9、icity,19/12/2020,Artificial Neural Networks - I,11,Binary Neurons,“Hard” threshold,= threshold,ex: Perceptrons, Hopfield NNs, Boltzmann Machines Main drawbacks: can only map binary functions, biologically implausible.,off,on,Stimulus,Response,19/12/2020,Artificial Neural Networks - I,12,Analog Neuro
10、ns,“Soft” threshold,ex: MLPs, Recurrent NNs, RBF NNs. Main drawbacks: difficult to process time patterns, biologically implausible.,off,on,Stimulus,Response,19/12/2020,Artificial Neural Networks - I,13,Spiking Neurons,= spike and afterspike potential urest = resting potential e(t,u(t) = trace at tim
11、e t of input at time t = threshold xj(t) = output of neuron j at time t wij = efficacy of synapse from neuron i to neuron j u(t) = input stimulus at time t,Response,Stimulus,19/12/2020,Artificial Neural Networks - I,14,Spiking Neuron Dynamics,19/12/2020,Artificial Neural Networks - I,15,赫布律,加拿大心理学家D
12、onald Hebb出版了行为的组织一书,指出学习导致突触的联系强度和传递效能的提高,即为“赫布律”。 在此基础上,人们提出了各种学习规则和算法,以适应不同网络模型的需要。有效的学习算法,使得神经网络能够通过连接权值的调整,构造客观世界的内在表示,形成具有特色的信息处理方法,信息存储和处理体现在网络的连接中。,19/12/2020,Artificial Neural Networks - I,16,Hebbs Postulate of Learning,Biological formulation When an axon of cell A is near enough to excite
13、a cell and repeatedly or persistently takes part in firing it, some growth process or metabolic change takes place in one or both cells such that As efficiency as one of the cells firing B is increased.,19/12/2020,Artificial Neural Networks - I,17,赫布律,当细胞A的一个轴突和细胞B 很近,足以对它产生影响,并且持久地、不断地参与了对细胞B 的兴奋,那
14、么在这两个细胞或其中之一会发生某种生长过程或新陈代谢变化,以致于A作为能使B 兴奋的细胞之一,它的影响加强了。,19/12/2020,Artificial Neural Networks - I,18,Hebbs Postulate: revisited,Stent (1973), and Changeux and Danchin (1976) have expanded Hebbs rule such that it also mo- dels inhibitory synapses: If two neurons on either side of a synapse are activa
15、ted simultaneously (synchronously), then the strength of that synapse is selectively increased. If two neurons on either side of a synapse are activated asynchronously, then that synapse is selectively weakened or eliminated.,19/12/2020,Artificial Neural Networks - I,19,Artificial Neural Networks,Ou
16、tput layer,Input layer,Hidden layers,fully connected,sparsely connected,19/12/2020,Artificial Neural Networks - I,20,Feedforward ANN Architectures,Information flow unidirectional Static mapping: y=f(x) Multi-Layer Perceptron (MLP) Radial Basis Function (RBF) Kohonen Self-Organising Map (SOM),19/12/2
17、020,Artificial Neural Networks - I,21,Recurrent ANN Architectures,Feedback connections Dynamic memory: y(t+1)=f(x(),y(),s() (t,t-1,.) Jordan/Elman ANNs Hopfield Adaptive Resonance Theory (ART),19/12/2020,Artificial Neural Networks - I,22,History,Early stages 1943 McCulloch-Pitts: neuron as comp. ele
18、m. 1948 Wiener: cybernatics 1949 Hebb: learning rule 1958 Rosenblatt: perceptron 1960 Widrow-Hoff: least mean square algorithm Recession 1969 Minsky-Papert: limitations perceptron model Revival 1982 Hopfield: recurrent network model 1982 Kohonen: self-organizing maps 1986 Rumelhart et. al.: backprop
19、agation,19/12/2020,Artificial Neural Networks - I,23,历史,40年代心理学家Mcculloch和数学家Pitts合作提出的兴奋与抑制型神经元模型和Hebb提出的神经元连接强度的修改规则,他们的研究结果至今仍是许多神经网络模型研究的基础。 50年代、60年代的代表性工作是Rosenblatt的感知机和Widrow的自适应性元件Adaline。 1969年,Minsky和Papert合作发表了颇有影响的Perceptron一书,得出了消极悲观的论点,加上数字计算机正处于全盛时期并在人工智能领域取得显著成就,70年代人工神经网络的研究处于低潮。,1
20、9/12/2020,Artificial Neural Networks - I,24,历史,80年代后,传统的Von Neumann数字计算机在模拟视听觉的人工智能方面遇到了物理上不可逾越的极限。与此同时,Rumelhart与Mcclelland以及Hopfield等人在神经网络领域取得了突破性进展,神经网络的热潮再次掀起。 自适应共振理论(ART) 组织特征映射理论 Hinton 等人最近提出了 Helmboltz 机 徐雷提出的 Ying-Yang 机理论模型 甘利俊一( S.Amari) 开创和发展的基于统计流形的方法应用于人工神经网络的研究,19/12/2020,Artificial
21、 Neural Networks - I,25,ANN Capabilities,Learning Approximate reasoning Generalisation capability Noise filtering Parallel processing Distributed knowledge base Fault tolerance,19/12/2020,Artificial Neural Networks - I,26,Main Problems with ANN,Knowledge base not transparent (black box) (Partially r
22、esolved) Learning sometimes difficult/slow Limited storage capability,19/12/2020,Artificial Neural Networks - I,27,ANN Learning Paradigms,Supervised learning Classification Control Function approximation Associative memory Unsupervised learning Clustering Reinforcement learning Control,19/12/2020,Ar
23、tificial Neural Networks - I,28,Supervised Learning,Teacher presents ANN input-output pairs ANN weights adjusted according to error Iterative algorithms (e.g. Delta rule, BP rule) One-shot learning (Hopfield) Quality of training examples is critical,19/12/2020,Artificial Neural Networks - I,29,Prese
24、nted by Martin Ho, Eddy Li, Eric Wong and Kitty Wong - Copyright 2000,Linear Separability in Perceptrons,19/12/2020,Artificial Neural Networks - I,30,Presented by Martin Ho, Eddy Li, Eric Wong and Kitty Wong - Copyright 2000,Learning Linearly Separable Functions (1),What can these functions learn ?
25、Bad news: - There are not many linearly separable functions. Good news: - There is a perceptron algorithm that will learn any linearly separable function, given enough training examples.,19/12/2020,Artificial Neural Networks - I,31,Delta Rule,a.k.a. Least Mean Squares Widrow-Hoff iterative delta rul
26、e (1960) Gradient descent of the error surface Guaranteed to find minimum error configuration in single layer ANNs Stochastic approximation of desired behaviour,19/12/2020,Artificial Neural Networks - I,32,Unsupervised Learning,ANN adapts weights to cluster input data Hebbian learning Connection sti
27、mulus-response strengthened (hebbian) Competitive learning algorithms Kohonen zt is random measurement noise. Objective: find a neural network that represents the input / output transformation (a function) F(x,W) such that F(x,W) approximates g(x) for every x,19/12/2020,Artificial Neural Networks -
28、I,41,Learning to approximate,c is the learning parameter (usually a constant),19/12/2020,Artificial Neural Networks - I,42,Learning with a perceptron,Perceptron:,Data:,Error:,Learning:,A perceptron is able to learn a linear function.,19/12/2020,Artificial Neural Networks - I,43,Learning with RBF neu
29、ral networks,Only the synaptic weights of the output neuron are modified. An RBF neural network learns a nonlinear function.,19/12/2020,Artificial Neural Networks - I,44,Learning with MLP neural networks,MLP neural network: with p layers,Data:,Error:,x,yout,1 2 p-1 p,19/12/2020,Artificial Neural Net
30、works - I,45,Learning with backpropagation,Learning: Apply the chain rule for differentiation: calculate first the changes for the synaptic weights of the output neuron; calculate the changes backward starting from layer p-1, and propagate backward the local error terms.,The method is still relative
31、ly complicated but it is much simpler than the original optimisation problem.,19/12/2020,Artificial Neural Networks - I,46,Learning with general optimization,In general it is enough to have a single layer of nonlinear neurons in a neural network in order to learn to approximate a nonlinear function.
32、 In such case general optimisation may be applied without too much difficulty.,Example: an MLP neural network with a single hidden layer:,19/12/2020,Artificial Neural Networks - I,47,Learning with general optimization,19/12/2020,Artificial Neural Networks - I,48,New methods for learning with neural
33、networks,Bayesian learning: the distribution of the neural network parameters is learnt Support vector learning: the minimal representative subset of the available data is used to calculate the synaptic weights of the neurons,19/12/2020,Artificial Neural Networks - I,49,Reinforcement Learning,Sequen
34、tial tasks Desired action may not be known Critic evaluation of ANN behaviour Weights adjusted according to critic May require credit assignment Population-based learning Evolutionary Algorithms Swarming Techniques Immune Networks,19/12/2020,Artificial Neural Networks - I,50,ANN Summary,19/12/2020,A
35、rtificial Neural Networks - I,51,神经网络的集成,1996年,Sollich和Krogh 将神经网络集成定义为:“神经网络集成是用有限个神经网络对同一个问题进行学习,集成在某输入示例下的输出由构成集成的各神经网络在该示例下的输出共同决定”。,19/12/2020,Artificial Neural Networks - I,52,ANN Application Areas,Classification Clustering Associative memory Control Function approximation,19/12/2020,Artificia
36、l Neural Networks - I,53,ANN Classifier systems,Learning capability Statistical classifier systems Data driven Generalisation capability Handle and filter large input data Reconstruct noisy and incomplete patterns Classification rules not transparent,19/12/2020,Artificial Neural Networks - I,54,Appl
37、ications for ANN Classifiers,Pattern recognition Industrial inspection Fault diagnosis Image recognition Target recognition Speech recognition Natural language processing Character recognition Handwriting recognition Automatic text-to-speech conversion,19/12/2020,Artificial Neural Networks - I,55,Cl
38、ustering with ANNs,Fast parallel distributed processing Handle large input information Robust to noise and incomplete patterns Data driven Plasticity/Adaptation Visualisation of results Accuracy sometimes poor,19/12/2020,Artificial Neural Networks - I,56,ANN Clustering Applications,Natural language
39、processing Document clustering Document retrieval Automatic query Image segmentation Data mining Data set partitioning Detection of emerging clusters Fuzzy partitioning Condition-action association,19/12/2020,Artificial Neural Networks - I,57,Associative ANN Memories,Stimulus-response association Au
40、to-associative memory Content addressable memory Fast parallel distributed processing Robust to noise and incomplete patterns Limited storage capability,19/12/2020,Artificial Neural Networks - I,58,Application of ANN Associative Memories,Character recognition Handwriting recognition Noise filtering
41、Data compression Information retrieval,19/12/2020,Artificial Neural Networks - I,59,ANN Control Systems,Learning/adaptation capability Data driven Non-linear mapping Fast response Fault tolerance Generalisation capability Handle and filter large input data Reconstruct noisy and incomplete patterns C
42、ontrol rules not transparent Learning may be problematic,19/12/2020,Artificial Neural Networks - I,60,ANN Control Schemes,ANN controller conventional controller + ANN for unknown or non-linear dynamics Indirect control schemes ANN models direct plant dynamics ANN models inverse plant dynamics,19/12/
43、2020,Artificial Neural Networks - I,61,ANN Control Applications,Non-linear process control Chemical reaction control Industrial process control Water treatment Intensive care of patients Servo control Robot manipulators Autonomous vehicles Automotive control Dynamic system control Helicopter flight
44、control Underwater robot control,19/12/2020,Artificial Neural Networks - I,62,ANN Function Modelling,ANN as universal function approximator Dynamic system modelling Learning capability Data driven Non-linear mapping Generalisation capability Handle and filter large input data Reconstruct noisy and i
45、ncomplete inputs,19/12/2020,Artificial Neural Networks - I,63,ANN Modelling Applications,Modelling of highly nonlinear industrial processes Financial market prediction Weather forecasts River flow prediction Fault/breakage prediction Monitoring of critically ill patients,19/12/2020,Artificial Neural Networks - I,64,Presented by Martin Ho, Eddy Li, Eric Wong and Kitty Wong - Cop
温馨提示
- 1. 本站所有资源如无特殊说明,都需要本地电脑安装OFFICE2007和PDF阅读器。图纸软件为CAD,CAXA,PROE,UG,SolidWorks等.压缩文件请下载最新的WinRAR软件解压。
- 2. 本站的文档不包含任何第三方提供的附件图纸等,如果需要附件,请联系上传者。文件的所有权益归上传用户所有。
- 3. 本站RAR压缩包中若带图纸,网页内容里面会有图纸预览,若没有图纸预览就没有图纸。
- 4. 未经权益所有人同意不得将文件中的内容挪作商业或盈利用途。
- 5. 人人文库网仅提供信息存储空间,仅对用户上传内容的表现方式做保护处理,对用户上传分享的文档内容本身不做任何修改或编辑,并不能对任何下载内容负责。
- 6. 下载文件中如有侵权或不适当内容,请与我们联系,我们立即纠正。
- 7. 本站不保证下载资源的准确性、安全性和完整性, 同时也不承担用户因使用这些下载资源对自己和他人造成任何形式的伤害或损失。
最新文档
- 2020吊篮操作证考试押题卷及答案通过率99%
- 2024物流专员笔试考前必背50题带精准答案
- 2023三资会计考试考前必背考题及答案完整版
- 2023青岛68中自招面试备考必看题库及答案完整版
- 2021初二微机会考全套模拟卷含完整答案解析
- 2025至2030城市建设规划产业绿色低碳发展路径与投资潜力评估
- 2026年配送中心管理考试试题及答案
- 2025-2030中国智慧农业技术应用市场调研与发展策略研究报告
- 2025-2030中国无人机物流配送试点效果评估与空域管理政策研究
- 2025-2030中国折叠屏手机面板良率提升与铰链技术专利壁垒分析
- 首件确认管理办法
- Q-JJJ 9002-2025 铁路建设项目安全穿透式管理实施指南
- 公共区域活动管理办法
- 护理岗位职责及工作流程
- 高三二轮复习生物种群群落生态系统微专题课件
- 内蒙古鄂尔多斯市基础建设有限公司招聘笔试题库2025
- 2025年中考数学压轴专题汇编(江苏专用)压轴专题09定角定高模型(原卷版+解析)
- 2024年江苏省高中学生英语口语等级测试试卷(模拟试卷)
- 教学课件-积极心理学(第2版)刘翔平
- 包钢集团笔试题库2025
- 2025党支部班子成员问题清单及整改措施
评论
0/150
提交评论