

新智元推荐
作者:RobbieAllen
整理:Sanglei,Shengsheng

去年,我写了一份相当受欢迎的博文(在Medium上有16万阅读量,见相关资源1),列出了我在深入研究大量机器学习资源时发现的最佳教程。十三个月后,现在有许多关于传统机器学习概念的新教程大量涌现以及过去一年中出现的新技术。围绕机器学习持续增加的大量内容有着惊人的数量。
本文包含了迄今为止我发现的最好的一些教程内容。它绝不是网上每个机器学习相关教程的简单详尽列表(这个工作量无疑是十分巨大而又枯燥重复的),而是经过详细筛选后的结果。我的目标就是将我在机器学习和自然语言处理领域各个方面找到的我认为最好的教程整理出来。
在教程中,为了能够更好的让读者理解其中的概念,我将避免罗列书中每章的详细内容,而是总结一些概念性的介绍内容。为什么不直接去买本书?当你想要对某些特定的主题或者不同方面进行了初步了解时,我相信这些教程对你可能帮助更大。
本文中我将分四个主题进行整理:机器学习,自然语言处理,Python和数学。在每个主题中我将包含一个例子和多个资源。当然我不可能完全覆盖所有的主题啦。
如果你发现我在这里遗漏了好的教程资源,请联系告诉我。为了避免资源重复罗列,我在每个主题下只列出了5、6个教程。下面的每个链接都应该链接了和其他链接不同的资源,也会通过不同的方式(例如幻灯片代码段)或者不同的角度呈现出这些内容。
相关资源本文作者RobbieAllen是以为科技作者和创业者、并自学AI并成为博士生。曾整理许多广为流传的机器学习相关资源。
1.2017版教程资源Over150oftheBestMachineLearning,NLP,andPythonTutorialsI’veFound(150多个最好的与机器学习,自然语言处理和Python相关的教程)
英文:
中文翻译:
英文:
中文翻译:
3.CheatSheetofMachineLearningandPython(andMath)CheatSheets
(值得收藏的27个机器学习速查表)
英文:
目录
1.机器学习1.1激活函数与损失函数1.2偏差(bias)1.3感知机(perceptron)1.4回归(Regression)1.5梯度下降(GradientDescent)1.6生成学习(GenerativeLearning)1.7支持向量机(SupportVectorMachines)1.8反向传播(Backpropagation)1.9深度学习(DeepLearning)1.10优化与降维(OptimizationandDimensionalityReduction)1.11LongShortTermMemory(LSTM)1.12卷积神经网络ConvolutionalNeuralNetworks(CNNs)1.13循环神经网络RecurrentNeuralNets(RNNs)1.14强化学习ReinforcementLearning1.15生产对抗模型GenerativeAdversarialNetworks(GANs)1.16多任务学习Multi-taskLearning2.自然语言处理NLP2.1深度学习与自然语言处理DeepLearningandNLP2.2词向量WordVectors2.3编解码模型Encoder-Decoder3.Python3.1样例Examples3.2Scipyandnumpy教程3.3scikit-learn教程3.4Tensorflow教程3.5PyTorch教程4.数学基础教程4.1线性代数4.2概率论4.3微积分第一部分:机器学习
StartHerewithMachineLearning()
MachineLearningisFun!(/@ageitgey)
MachineLearningCrashCourse:PartI,PartII,PartIII(MachineLearningatBerkeley)
PartI
PartII
PartIII
AnIntroductiontoMachineLearningTheoryandItsApplications:AVisualTutorialwithExamples()
AGentleGuidetoMachineLearning()
WhichmachinelearningalgorithmshouldIuse?()
TheMachineLearningPrimer()
MachineLearningTutorialforBeginners(/kanncaa1)
1.1激活函数与损失函数Whatistheroleoftheactivationfunctioninaneuralnetwork?()
Comprehensivelistofactivationfunctionsinneuralnetworkswithpros/cons()
Activationfunctionsandit’stypes-Whichisbetter?()
1.2偏差(bias)RoleofBiasinNeuralNetworks()
Whatisbiasinartificialneuralnetwork?()
1.3感知机(perceptron)Single-layerNeuralNetworks(Perceptrons)()
FromPerceptronstoDeepNetworks()
1.4回归(Regression)Introductiontolinearregressionanalysis()
LinearRegression()
LogisticRegression()
1.5梯度下降(GradientDescent)1.6生成学习(GenerativeLearning)ApracticalexplanationofaNaiveBayesclassifier()
1.7支持向量机(SupportVectorMachines)AnintroductiontoSupportVectorMachines(SVM)()
1.8反向传播(Backpropagation)Yesyoushouldunderstandbackprop(/@karpathy)
Canyougiveavisualexplanationforthebackpropagationalgorithmforneuralnetworks?(/rasbt)
BackpropagationThroughTimeandVanishingGradients()
1.9深度学习(DeepLearning)AGuidetoDeepLearningbyYN²()
DeepLearningPapersReadingRoadmap(/floodsung)
ATutorialonDeepLearning()
WhatisDeepLearning?()
What’stheDifferenceBetweenArtificialIntelligence,MachineLearning,andDeepLearning?()
DeepLearning—TheStraightDope()
1.10优化与降维(OptimizationandDimensionalityReduction)SevenTechniquesforDataDimensionalityReduction()
1.11LongShortTermMemory(LSTM)AGentleIntroductiontoLongShort-TermMemoryNetworksbytheExperts()
1.12卷积神经网络ConvolutionalNeuralNetworks(CNNs)DeepLearningandConvolutionalNeuralNetworks(/@ageitgey)
1.13循环神经网络RecurrentNeuralNets(RNNs)1.14强化学习ReinforcementLearningSimpleBeginner’sguidetoReinforcementLearningitsimplementation()
ATutorialforReinforcementLearning()
1.15生成对抗模型GenerativeAdversarialNetworks(GANs)AdversarialMachineLearning()
What’saGenerativeAdversarialNetwork?()
AbusingGenerativeAdversarialNetworkstoMake8-bitPixelArt(/@ageitgey)
AnintroductiontoGenerativeAdversarialNetworks(withcodeinTensorFlow)()
GenerativeAdversarialNetworksforBeginners()
1.16多任务学习Multi-taskLearning第二部分:自然语言处理
NaturalLanguageProcessingisFun!(/@ageitgey)
APrimeronNeuralNetworkModelsforNaturalLanguageProcessing(YoavGoldberg)
TheDefinitiveGuidetoNaturalLanguageProcessing()
IntroductiontoNaturalLanguageProcessing()
NaturalLanguageProcessing(almost)fromScratch()
2.1深度学习与自然语言处理DeepLearningandNLPDeepLearningappliedtoNLP()
DeepLearningforNLP(withoutMagic)(RichardSocher)
Embed,encode,att,predict:Thenewdeeplearningformulaforstate-of-the-artNLPmodels()
UnderstandingNaturalLanguagewithDeepNeuralNetworksUsingTorch()
2.2词向量WordVectorsBagofWordsMeetsBagsofPopcorn()
OnwordembeddingsPartI,PartII,PartIII()
PartI:
PartII:
PartIII:
Theamazingpowerofwordvectors()
word2vecParameterLearningExplained()
2.3编解码模型Encoder-DecoderSequencetoSequenceModels()
SequencetoSequenceLearningwithNeuralNetworks(NIPS2014)
MachineLearningisFunPart5:LanguageTranslationwithDeepLearningandtheMagicofSequences(/@ageitgey)
HowtouseanEncoder-DecoderLSTMtoEchoSequencesofRandomIntegers()
tf-seq2seq()
第三部分:Python
MachineLearningCrashCourse()
AwesomeMachineLearning(/josephmisiti)
Anexamplemachinelearningnotebook()
MachineLearningwithPython()
3.1样例ExamplesMLfromScatch(/eriklindernoren)
PythonMachineLearning(2ndEd.)CodeRepository(/rasbt)
3.2Scipyandnumpy教程ScipyLectureNotes()
AnintroductiontoNumpyandScipy(UCSBCHE210D)
3.3scikit-learn教程scikit-learnClassificationAlgorithms(/mmmayo13)
scikit-learnTutorials()
Abridgedscikit-learnTutorials(/mmmayo13)
3.4Tensorflow教程TensorflowTutorials()
IntroductiontoTensorFlow—CPUvsGPU(/@erikhallstrm)
TensorFlow:Aprimer()
3.5PyTorch教程Tutorial:DeepLearninginPyTorch()
PyTorchExamples(/jcjohnson)
PyTorchTutorial(/MorvanZhou)
PyTorchTutorialforDeepLearningResearchers(/yunjey)
第四部分:数学基础知识
MathforMachineLearning()
MathforMachineLearning(UMIACSCMSC422)
4.1线性代数AnIntuitiveGuidetoLinearAlgebra()
AProgrammer’sIntuitionforMatrixMultiplication()
UnderstandingtheCrossProduct()
UnderstandingtheDotProduct()
LinearAlgebraforMachineLearning()
Linearalgebracheatsheetfordeeplearning()
4.2概率论UnderstandingBayesTheoremWithRatios()
ProbabilityTheoryReviewforMachineLearning(StanfordCS229)
ProbabilityTheory()
ProbabilityTheoryforMachineLearning()
4.3微积分HowToUnderstandDerivatives:TheQuotientRule,Exponents,andLogarithms()
HowToUnderstandDerivatives:TheProduct,PowerChainRules()
VectorCalculus:UnderstandingtheGradient()
CalculusOverview()
新智元AIWORLD2018大会【早鸟票】开售!
大会官网:
即日起到8月19日,新智元限量发售若干早鸟票,与全球AI领袖近距离交流,见证全球人工智能产业跨越发展。
免责声明:本文章如果文章侵权,请联系我们处理,本站仅提供信息存储空间服务如因作品内容、版权和其他问题请于本站联系