人工智能詞匯12.17

1、常用英語詞匯-andrew Ng課程intensity 強(qiáng)度Regression 回歸Loss function損失函數(shù) non-convex非凸函數(shù) neural network 神經(jīng)網(wǎng)絡(luò) supervised learning 監(jiān)督學(xué)習(xí) regression problem回歸問題處理的是連續(xù)的問題 classification problem 分類問題 discreet value 離散值support vector machines 支持向雖機(jī) learning theory 學(xué)習(xí)理論 learning algorithms 學(xué)習(xí)算法 unsupervised learning 無監(jiān)督

2、學(xué)習(xí) gradient descent 梯度卜降 linear regression 線性回歸 Neural Network 神經(jīng)網(wǎng)絡(luò) gradient descent 梯度卜降 normal equations linear algebra 線性代數(shù)superscript 上標(biāo) exponentiation 指數(shù) training set訓(xùn)練集合 training example 訓(xùn)練樣本 hypothesis假設(shè)，用來表示學(xué)習(xí)算法的輸岀 LMS algorithm least mean squares 最小二乘法算法 batch gradient descent 批量梯度下降 consta

3、ntly gradient descent 隨機(jī)梯度下降 iterative algorithm 迭代算法 partial derivative 偏導(dǎo)數(shù)contour等髙線 quadratic function 二元函數(shù) locally weighted regression 局部加權(quán)回歸 underfitting 欠擬合 overfitting 過擬合non-parametric learning algorithms 無參數(shù)學(xué)習(xí)算法 parametric learning algorithm 參數(shù)學(xué)習(xí)算法 activation 激活值activation function 激活函數(shù) add

4、itive noise 加性噪聲autoencoder自編碼器Autoencoders自編碼算法 average firing rate 平均激活率 average sumof-squares error 均方差 backpropagation 后向傳播 basis 基 basis feature vectors 特征基向量 batch gradient ascent批量梯度上升法Bayesian regularization method 貝葉斯規(guī)則化方法Bernoulli random variable 伯努利隨機(jī)變量bias term偏宜項(xiàng)binary classfication 二元分

5、類class labels類型標(biāo)記concatena廿on 級(jí)聯(lián)conjugate gradient 共$6梯度contiguous groups 聯(lián)通區(qū)域convex optimization software 凸優(yōu)化軟件 convolution 卷積cost function代價(jià)函數(shù) covariance matrix 協(xié)方差矩陣DC component直流分量decorrelation 去相關(guān)degeneracy 退化demensionality reduction 降維derivative 導(dǎo)函數(shù)diagonal對(duì)角線diffusion of gradients 梯度的彌散eigenv

6、alue 特征值 eigenvector特征向量 error term 殘差feature matrix 特征矩陣feature standardization 特征標(biāo)準(zhǔn)化 feedforward architectures 前饋結(jié)構(gòu)算法 feedforward neural network 前饋神經(jīng)網(wǎng)絡(luò) feedforward pass 前饋傳導(dǎo)fine-tuned 微調(diào)first-order feature 階特征forward pass前向傳導(dǎo)forward propagation 前向傳播Gaussian prior髙斯先驗(yàn)概率generative model 生成模型 gradien

7、t descent 梯度卜降 Greedy layer-wise training逐層貪婪訓(xùn)練方法 grouping matrix 分組矩陣 Hadamard product阿達(dá)馬乘積 Hessian matrix Hessian 矩陣 hidden layer 隱含層 hidden units隱藏神經(jīng)元 Hierarchical grouping 層次型分組 higher-order features 更高階特征highly non-convex optimization problem 高度非凸的優(yōu)化問題 histogram直方圖 hyperbolic tangent雙曲正切函數(shù) hypo

8、thesis估值，假設(shè) identity activation function 恒等激勵(lì)函數(shù) HD獨(dú)立同分布 illumination 照明 inactive 抑制 independent component analysis 獨(dú)立成份分析input domains 輸入域 input layer 輸入層 intensity亮度/灰度 intercept term 截距 KL divergence 相對(duì)爛 KL divergence KL 分散度 k-Means K均值 learning rate學(xué)習(xí)速率 least squares 最小二乘法 linear correspondence 線

9、性響應(yīng)linear superposition 線性疊加 line-search algorithm 線搜索算法 local mean subtraction 局部均值消減 local optima局部最優(yōu)解 logistic regression 邏輯回歸 loss function損失函數(shù) low-pass filtering 低通濾波 magnitude 幅值 MAP極大后驗(yàn)估計(jì) maximum likelihood estimation 極大似然估計(jì) mean平均值MFCC Mel倒頻系數(shù) multi-class classification 多元分類 neural networks

10、神經(jīng)網(wǎng)絡(luò) neuron神經(jīng)元Newtons method 牛頓法 non-convex function 非凸函數(shù) non-linear feature 非線性特征 norm范式 norm bounded有界范數(shù)norm constrained 范數(shù)約朿 normalization 歸一ft numerical roundoff errors 數(shù)值舍入誤差 numerically checking 數(shù)值檢驗(yàn) numerically reliable數(shù)值計(jì)算上穩(wěn)定 object detection 物體檢測(cè) objective function 目標(biāo)函數(shù) off-by-one error 缺位錯(cuò)

11、誤 orthogonalization 正交 output layer 輸出層overall cost function 總體代價(jià)函數(shù) over-complete basis 超完備基 over-fitting 過擬合 parts of objects目標(biāo)的部件 part-whole decompostion 部分整體分解 PCA主元分析 penalty term懲罰因子 per-example mean subtraction 逐樣本均值消減 pooling 池化 pretrain預(yù)訓(xùn)練principal components analysis 主成份分析 quadratic constra

12、ints 二次約束 RBMs 受限 Boltzman 機(jī) reconstruction based models 基于重構(gòu)的模型 reconstruction cost 重建代價(jià) reconstruction term 重*勾項(xiàng) redundant 冗余 reflection matrix 反射矩陣 regularization 正則化 regularization term 正則化項(xiàng) rescaling 縮放 robust魯棒性 run行程 second-order feature 二階特征 sigmoid activation function S 型激勵(lì)函數(shù) significant di

13、gits 有效數(shù)字 singular value 奇異值 singular vector 奇異向呈 smoothed LI penalty平滑的LI范數(shù)懲罰 Smoothed topographic LI sparsity penalty平滑地形LI稀疏懲罰函數(shù) smoothing 平滑Softmax Regress on Softmax 回歸 sorted in decreasing order 降序排列 source features 源特征 sparse autoencoder 消減歸一化 Sparsity稀疏性 sparsity parameter 稀疏性參數(shù) sparsity pen

14、alty 稀疏懲罰 square function 平方函數(shù)squared-error 方差 stationary平穩(wěn)性（不變性） stationary stochastic process 平穩(wěn)隨機(jī)過程 step-size步長(zhǎng)值 supervised learning 監(jiān)督學(xué)習(xí)symmetric positive semi-definite matrix對(duì)稱半正定矩陣symmetry breaking 對(duì)稱失效 tanh function雙曲正切函數(shù) the average activation 平均活躍度the derivative checking method 梯度驗(yàn)證方法 the e

15、mpirical distribution 經(jīng)驗(yàn)分布函數(shù) the energy function 能量函數(shù) the Lagrange dual拉格朗日對(duì)偶函數(shù) the log likelihood對(duì)數(shù)似然函數(shù) the pixel intensity value 像素灰度值 the rate of convergence 收斂速度 topographic cost term 拓?fù)浯鷥r(jià)項(xiàng) topographic ordered 拓扌卜秩序 transformation 變換 translation invariant 平移不變性 trivial answer 平凡解 under-complete

16、basis 不完備基 unrolling組合擴(kuò)展 unsupervised learning 無監(jiān)督學(xué)習(xí) variance 方差 vecotrized implementation 向量化實(shí)現(xiàn) vectorization 矢量化 visual cortex視覺皮層 weight decay權(quán)重衰減weighted average加權(quán)平均值 whitening 白化 zero-mean均值為零 Accumulated error backpropagation 累積誤差逆?zhèn)鞑?Activation Function 激活函數(shù)Adaptive Resonance Theory/ART 自適應(yīng)諧振理

17、論 Addictive model 加性學(xué)習(xí)Adversarial Networks 對(duì)抗網(wǎng)絡(luò)Affine Layer 仿射層 Affinity matrix 親和矩陣Agent代理/智能體Algorithm 算法 Alpha-beta pruning a-p 剪枝 Anomaly detection 異常檢測(cè) Approximation 近似Area Under ROC Curve / AUC Roc 曲線卜而積 Artificial General Intelligence/AGI 通用人匸智能 Artificial Intelligence/AI 人工智能 Association ana

18、lysis 關(guān)聯(lián)分析Attention mechanism 注意力機(jī)制Attribute condi廿onal independence assumption 屬性條件獨(dú)立性假設(shè)Attribute space 屬性空間 Attribute value 屬性值 Autoencoder自編碼器 Automatic speech recognition 自動(dòng)語音識(shí)別 Automatic summarization 自動(dòng)摘要 Average gradient 平均梯度 Average-Pooling 平均池化Backpropagation Through Time通過時(shí)間的反向傳播Backpropag

19、ation/BP 反向傳播Base learner基學(xué)習(xí)器Base learning algorithm 基學(xué)習(xí)算法Batch Normalization/BN 批量歸一化Bayes decision rule貝葉斯判定準(zhǔn)貝!IBayes Model Averaging / BMA 貝葉斯模型平均Bayes optimal classifier貝葉斯最優(yōu)分類器Bayesian decision theory 貝葉斯決策論Bayesian network貝葉斯網(wǎng)絡(luò)Between-class scatter matrix 類間散度矩陣Bias偏置/偏差Bias-variance decomposi

20、tion 偏差方差分解Bias-Variance Dilemma 偏差-方差困境Bi-directional Long-Short Term Memory/Bi-LSTM雙向長(zhǎng)短期記憶Binary classification 二分類Binomial test 二項(xiàng)檢驗(yàn)Bi-partition 二分法Boltzmannmachine 玻爾茲曼機(jī)Bootstrap sampling自助采樣法/可重復(fù)采樣Bootstrapping 自助法Break-Event Point / BEP 平衡點(diǎn)Calibration 校準(zhǔn)Cascade-Correlation 級(jí)聯(lián)相關(guān)Categorical attr

21、ibute 離散屬性Class-conditional probability 類條件概率Classification and regression tree/CART 分類打回歸樹Classifier 分類器Class-imbalance類別不平衡Closed -form 閉式Cluster簇/類/集群Cluster analysis 聚類分析Clustering 聚類Clustering ensemble 聚類集成Co-adapting 共適應(yīng)Coding matrix編碼矩陣COLT國(guó)際學(xué)習(xí)理論會(huì)議Committee-based learning基于委員會(huì)的學(xué)習(xí) Competitive

22、learning 競(jìng)爭(zhēng)型學(xué)習(xí)Component learner 組件學(xué)習(xí)器Comprehensibility 可解釋性Computation Cost 訂算成本Computational Linguistics IT算語言學(xué)Computer vision計(jì)算機(jī)視覺Concept drift概念漂移Concept Learning System /CLS 概念學(xué)習(xí)系統(tǒng)Conditional entropy 條件爛Conditional mutual information 條件互信息Conditional Probability Table / CPT 條件概率表Conditional rand

23、om field/CRF 條件隨機(jī)場(chǎng)Conditional risk 條件風(fēng)險(xiǎn)Confidence 置信度Confusion matrix 混淆矩陣Connection weight 連接權(quán)Connectionism 連結(jié)主義Consistency 一致性/相合性Contingency table 列聯(lián)表Continuous attribute 連續(xù)屬性Convergence 收斂Conversational agent 會(huì)話智能體Convex quadratic programming 凸二次規(guī)劃 Convexity 凸性Convolutional neural network/CNN 卷積

24、神經(jīng)網(wǎng)絡(luò) Cooccurrenee 同現(xiàn)Correlation coefficient 相關(guān)系數(shù)Cosine similarity 余弦相似度Cost curve成本曲線Cost Function成本函數(shù)Cost matrix成本矩陣Cost-sensitive 成本敏感Cross entropy 交叉爛Cross validation 交叉驗(yàn)證Crowdsourcing 眾包Curse of dimensionality 維數(shù)災(zāi)難Cut point截?cái)帱c(diǎn)Cutting plane algorithm 割平面法Data mining數(shù)據(jù)挖掘Data set數(shù)據(jù)集Decision Boundar

25、y 決策邊界Decision stump 決策樹樁Decision tree決策樹/判定樹Deduction 演繹Deep Belief Network深度信念網(wǎng)絡(luò)Deep Convolutional Generative Adversarial NetworkDCGAN深度卷積生成對(duì)抗網(wǎng)絡(luò)Deep learning深度學(xué)習(xí)Deep neural network/DNN 深度神經(jīng)網(wǎng)絡(luò)Deep Q-Learning 深度 Q 學(xué)習(xí)Deep Q-Network 深度 Q 網(wǎng)絡(luò)Density estimation 密度估訃Density-based clustering 密度聚類Differenti

26、able neural computer 可微分神經(jīng)計(jì)算機(jī) Dimensionality reduction algorithm 降維算法 Directed edge 有向邊Disagreement measure 不合度量Discriminative model 判別模型Discriminator 判別器Distance measure 距離度量Distance metric learning 距離度量學(xué)習(xí)Distribution 分布Divergence 散度Diversity measure多樣性度量/差異性度量 Domain adaption領(lǐng)域自適應(yīng)Downsampling 卜采樣D

27、-separation (Directed separation) 有向分離 Dual problem對(duì)偶問題Dummy node啞結(jié)點(diǎn)Dynamic Fusion 動(dòng)態(tài)融合Dynamic programming 動(dòng)態(tài)規(guī)劃Eigenvalue decomposition 特征值分解Embedding 嵌入Emotional analysis 情緒分析Empirical conditional entropy 經(jīng)驗(yàn)條件埼Empirical entropy 經(jīng)驗(yàn)埼Empirical error 經(jīng)驗(yàn)誤差Empirical risk 經(jīng)驗(yàn)風(fēng)險(xiǎn)End-to-End 端到端Energy-based mo

28、del基于能雖:的模型Ensemble learning 集成學(xué)習(xí)Ensemble pruning 集成修剪Error Correcting Output Codes / ECOC 糾錯(cuò)輸出碼 Error rate錯(cuò)誤率Error-ambiguity decomposition 誤差分歧分解 Euclidean distance 歐氏距離Evolutionarycomputa廿on i寅算 Expectation-Maximization 期望最大化Expected loss期望損失Exploding Gradient Problem 梯度爆炸問題 Exponential loss funct

29、ion 指數(shù)損失函數(shù)Extreme Learning Machine/ELM 超限學(xué)習(xí)機(jī)Factorization 因子分解False negative 假負(fù)類False positive 假正類False Positive Rate/FPR 假正例率Feature engineering 特征工程Feature selection 特征選擇Feature vector 特征向量Featured Learning 特征學(xué)習(xí)Feedforward Neural Networks/FNN 前饋神經(jīng)網(wǎng)絡(luò) Fine-tuning 微調(diào)Flipping output 翻轉(zhuǎn)法Fluctuation 喪蕩F

30、orward stagewise algorithm 前向分步算法Frequentist頻率主義學(xué)派Full-rank matrix 滿秩矩陣Functional neuron 功能神經(jīng)元Gain ratio增益率Game theory 博弈論Gaussian kernel function 高斯核函數(shù)Gaussian Mixture Model 髙斯混合模型General Problem Solving 通用問題求解Generalization 泛化Generalization error 泛化誤差Generalization error bound 泛化誤差上界 Generalized L

31、agrange function 廣義拉格朗日函數(shù) Generalized linear model / 義線性模型 Generalized Rayleigh quotient 丿義瑞利商 Generative Adversarial Networks/GAN 生成對(duì)抗網(wǎng)絡(luò) Generative Model 生成模型Generator生成器Genetic Algorithm/GA 遺傳算法Gibbs sampling吉布斯采樣Gini index基尼指數(shù)Global minimum 全局最小Global Optimization 全局優(yōu)化Gradient boosting 梯度提升Gradie

32、nt Descent 梯度下降Graph theory 圖論Ground-truth M相 / 真實(shí)Hard margin 硬間隔Hard voting 硬投票Harmonic mean調(diào)和平均Hesse matrix海塞矩陣Hidden dynamic model 隱動(dòng)態(tài)模型Hidden layer 隱藏層Hidden Markov Model/HMM隱馬爾可夫模型Hierarchical clustering 層次聚類Hilbert space希爾伯特空間Hinge loss function合頁損失函數(shù)Hold-out留岀法Homogeneous 同質(zhì)Hybrid computing 混

33、合計(jì)算Hyperparameter 超參數(shù)Hypothesis 假設(shè)Hypothesis test 假設(shè)驗(yàn)證ICML國(guó)際機(jī)器學(xué)習(xí)會(huì)議Improved iterative scaling/IIS 改進(jìn)的迭代尺度法Incremental learning 增戢學(xué)習(xí)Independent and identically distributed/ 獨(dú)立同分 布Independent Component Analysis/ICA 獨(dú)立成分分析Indicator function 指示函數(shù)Individual learner 個(gè)體學(xué)習(xí)器Induction 歸納Inductive bias 歸納偏好Indu

34、ctive learning 歸納學(xué)習(xí)Inductive Logic Programming / ILP 歸納邏借程序設(shè)計(jì)Information entropy 信息矯Information gain 信息增益Input layer 輸入層Insensitive loss不敏感損失Inter-cluster similarity 簇間相似度International Con fere nee for Machine Learning/ICML國(guó)際機(jī)器學(xué)習(xí)大會(huì)Intra-cluster similarity 簇內(nèi)相似度Intrinsic value 固有值Isometric Mapping/ls

35、omap 等度量映射Isotonic regression 等分回歸Iterative Dichotomiser 迭代二分器Kernel method 核方法Kernel trick 核技巧Kernelized Linear Discriminant Analysis / KLDA核線性判別分析K-fold cross validation k折交叉驗(yàn)證/ k倍交叉驗(yàn)證K-Means Clustering K -均值聚類K-Nearest Neighbours Algorithm/KNN K 近鄰算法Knowledge base 知識(shí)庫Knowledge Representation 知識(shí)表征

36、Label space標(biāo)記空間Lagrange duality拉格朗日對(duì)偶性Lagrange multiplier拉格朗日乘子Laplace smoothing拉普拉斯平滑Laplacian correction 拉普拉斯修正Latent Dirichlet Allocation 隱狄利克雷分布Latent semantic analysis 潛在語義分析Latent variable 隱變量Lazy learning懶惰學(xué)習(xí)Learner學(xué)習(xí)器Learning by analogy 類比學(xué)習(xí)Learning rate 學(xué)習(xí)率Learning Vector Quantization/LVQ 學(xué)

37、習(xí)向量量化Least squares regression tree 最小二乘回歸樹Leave-One-Out/LOO 留一法linear chain conditional random field線性鏈條件隨機(jī)場(chǎng)Linear Discriminant Analysis / LDA 線性判別分析Linear model線性模型Linear Regression 線性回歸Link function聯(lián)系函數(shù)Local Markov property局部馬爾可夫性Local minimum局部最小Log likelihood 對(duì)數(shù)似然Log odds / logit 對(duì)數(shù)幾率Logis廿c Reg

38、ression Logistic 回歸Log-likelihood 對(duì)數(shù)似然Log-linear regression 對(duì)數(shù)線性回歸Long-Short Term Memory/LSTM 長(zhǎng)短期記憶Loss function損失函數(shù)Machine translation/MT 機(jī)器翻譯Macron-P宏查準(zhǔn)率Macron-R宏查全率Majority voting絕對(duì)多數(shù)投票法Manifold assumption 流形假設(shè)Manifold learning 流形學(xué)習(xí)Margin theory間隔理論Marginal distribution 邊際分布Marginal independence

39、邊際獨(dú)立性Marginalization 邊際化Markov Chain Monte Carlo/MCMC馬爾可夫鏈蒙特卡羅方法Markov Random Field馬爾可夫隨機(jī)場(chǎng)Maximal clique 最大團(tuán)Maximum Likelihood Estimation/MLE極大似然估計(jì)/極大似然法Maximum margin 最大間隔Maximum weighted spanning tree 最大帶權(quán)生成樹Max-Pooling最大池化Mean squared error 均方誤差Meta-learner元學(xué)習(xí)器Metric learning 度量學(xué)習(xí)Micro-P微査準(zhǔn)率Micro

40、-R微査全率Minimal Description Length/MDL 最小描述長(zhǎng)度Minimax game極小極大博弈Misclassification cost 誤分類成本Mixture of experts 混合&家Momentum 動(dòng)量Moral graph道徳圖/端正圖Multi-class classification 多分類Multi-document summarization 多文檔摘要Multi-layer feedforward neural networks多層前饋神經(jīng)網(wǎng)絡(luò)Multilayer Perceptron/MLP 多層感知器Multimodal learni

41、ng 多模態(tài)學(xué)習(xí)Multiple Dimensional Scaling 多維縮放Multiple linear regression 多元線性回歸Multi-response Linear Regression / MLR多響應(yīng)線性回歸Mutual information 互信息Naive bayes樸素貝葉斯Naive Bayes Classifier樸素貝葉斯分類器Named entity recognition 命 名實(shí)體識(shí)別Nash equilibrium 納什均衡Natural language generatlon/NLG 自然語言生成Natural language proce

42、ssing 自然語言處理Negative class 負(fù)類Negative correlation 負(fù)相關(guān)法Negative Log Likelihood 負(fù)對(duì)數(shù)似然Neighbourhood Component Analysis/NCA近鄰成分分析Neural Machine Translation 神經(jīng)機(jī)器翻譯Neural Turing Machine 神經(jīng)圖靈機(jī)Newton method 牛頓法NIPS國(guó)際神經(jīng)信息處理系統(tǒng)會(huì)議No Free Lunch Theorem / NFL沒有免費(fèi)的午餐左理Noise-contrastive estimation 噪音對(duì)比估計(jì)Nominal att

43、ribute 列名屬性Non-convex optimization 非凸優(yōu)化Nonlinear model非線性模型Non-metric distance 非度量距離Non-negative matrix factorization 非負(fù)矩陣分解Non-ordinal attribute 無序?qū)傩訬on-Saturating Game 非飽和博弈Norm范數(shù)Normalization 歸一化Nuclear norm 核范數(shù)Numerical attribute 數(shù)值屬性Letter 0Objective function 目標(biāo)函數(shù)Oblique decision tree 斜決策樹Occam

44、s razor奧卡姆剃刀Odds幾率Off-Policy 離策略O(shè)ne shot learning 一次性學(xué)習(xí)One-Dependent Estimator / ODE 獨(dú)依賴估計(jì)On-Policy在策略O(shè)rdinal attribute 有序?qū)傩設(shè)ut-of-bag estimate 包外估計(jì)Output layer 輸出層Output smearing輸出調(diào)制法Overfitting過擬合/過配Oversampling 過采樣Paired t-test 成對(duì) t 檢驗(yàn)Pairwise成對(duì)型Pairwise Markov property 成對(duì)馬爾可夫性Parameter 參數(shù)Paramet

45、er estimation 參數(shù)估計(jì)Parameter tuning 調(diào)參Parse tree解析樹Particle Swarm Optimization/PSO 粒子群優(yōu)化算法Part-of-speech tagging 詞性標(biāo)注Perceptron 感知機(jī)Performance measure 性能度量Plug and Play Generative Network 即插即用生成網(wǎng)絡(luò)Plurality voting相對(duì)多數(shù)投票法Polarity detection 極性檢測(cè)Polynomial kernel function 多項(xiàng)式核函數(shù)Pooling 池化Positive class 正

46、類Positive definite matrix 正圧矩陣Post-hoc test后續(xù)檢驗(yàn)Post-pruning 后剪枝potential function 勢(shì)函數(shù)Precision查準(zhǔn)率/準(zhǔn)確率Prepruning 預(yù)剪枝Principal component analysis/PCA 主成分分析Principle of multiple explanations 多釋原則Prior先驗(yàn)Probability Graphical Model 概率圖模型Proximal Gradient Descent/PGD 近端梯度下降Pruning 剪枝Pseudo-label 偽標(biāo)記Quanti

47、zed Neural Network 量子化神經(jīng)網(wǎng)絡(luò)Quantum computer量子汁算機(jī)Quantum Computing 量子計(jì)算Quasi Newton method 擬牛頓法Radial Basis Function / RBF 徑向基函數(shù)Random Forest Algorithm 隨機(jī)森林算法Random walk隨機(jī)漫步Recall查全率/召回率Receiver Operating Characteristic/ROC受試者工作特征Rectified Linear Unit/ReLU 線性修正單元Recurrent Neural Network 循環(huán)神經(jīng)網(wǎng)絡(luò)Recursiv

48、e neural network 遞歸神經(jīng)網(wǎng)絡(luò)Reference model 參考模型Regression 回歸Regularization 正則化Reinforcement learning/RL 強(qiáng)化學(xué)習(xí)Representation learning 表征學(xué)習(xí)Representer theorem 表示左理 reproducing kernel Hilbert space/RKHS再生核希爾伯特空間Re-sampling重釆樣法Rescaling再縮放Residual Mapping 殘差映射Residual Network 殘差網(wǎng)絡(luò)Restricted Boltzmann Machine

49、/RBM 受限玻爾茲曼機(jī)Restricted Isometry Property/RIP 限定等距性 Re-weighting重賦權(quán)法Robustness穩(wěn)健性/魯棒性Root node根結(jié)點(diǎn)Rule Engine規(guī)則引擎Rule learning規(guī)則學(xué)習(xí)Saddle point 鞍點(diǎn)Sample space樣本空間Sampling 采樣Score function 評(píng)分函數(shù)Self-Driving自動(dòng)駕駛Self-Organizing Map / SOM 自組織映射Semi-naive Bayes classifiers半樸素貝葉斯分類器 Semi-Supervised Learning 半

50、監(jiān)督學(xué)習(xí) semi-Supervised Support Vector Machine半監(jiān)督支持向量機(jī)Sentiment analysis 情感分析Separating hyperplane 分離超平而Sigmoid funotion Sigmoid 函數(shù)Similarity measure 相似度度量Simulated annealing 模擬退火Simultaneous localization and mapping同步定位與地圖構(gòu)建Singular Value Decomposition 奇異值分解Slack variables 松弛變量Smoothing 平滑Soft margin

51、軟間隔Soft margin maximization 軟間隔最大化Soft voting 軟投票Sparse representation 稀疏表征Sparsity稀疏性Specialization 特化Spectral Clustering 譜聚類Speech Recognition 語音識(shí)別Splitting variable 切分變量Squashing function 擠壓函數(shù)Stability-plasticity dilemma 可塑性穩(wěn)宦性困境Statistical learning 統(tǒng)計(jì)學(xué)習(xí)Status feature function 狀態(tài)特征函Stochastic gra

52、dient descent 隨機(jī)梯度下降Stratified sampling 分層采樣Structural risk 結(jié)構(gòu)風(fēng)險(xiǎn)Structural risk minimization/SRM 結(jié)構(gòu)風(fēng)險(xiǎn)最小化Subspace子空間Supervised learning監(jiān)督學(xué)習(xí)/有導(dǎo)師學(xué)習(xí) support vector expansion 支持向量展式 Support Vector Machine/SVM 支持向量機(jī) Surrogat loss替代損失Surrogate function 替代函數(shù)Symbolic learning 符號(hào)學(xué)習(xí)Symbolism符號(hào)主義Synset同義詞集T-Dis

53、tribution Stochastic Neighbour Embedding t-SNET-分布隨機(jī)近鄰嵌入Tensor張量Tensor Processing Units/TPU 張量處理單元The least square method 最小二乘法Threshold 閾值Threshold logic unit閾值邏輯單元 Threshold-moving 閾值移動(dòng) Time Step時(shí)間步驟Tokenization 標(biāo)記化Training error 訓(xùn)練誤差Training instance訓(xùn)練示例/訓(xùn)練例Transductive learning 直推學(xué)習(xí)Transfer lear

54、ning 遷移學(xué)習(xí)Treebank 樹庫Tria-by-error 試錯(cuò)法True negative 頁負(fù)類True positive 真正類True Positive Rate/TPR 貞.正例率Turing Machine 圖靈機(jī)Twice-learning 二次學(xué)AlUnderfitting欠擬合/欠配Undersampling 欠采樣Understandability 可理解性Unequal cost非均等代價(jià)Unit-step function單位階躍函數(shù)Univariate decision tree 單變量決策樹Unsupervised learning無監(jiān)督學(xué)習(xí)/無導(dǎo)師學(xué)習(xí)Un

55、supervised layer-wise training 無監(jiān)督逐層訓(xùn)練 Upsampling 上采樣Vanishing Gradient Problem 梯度消失問題 Variational inference 變分推斷VC Theory VC 維理論Version space版本空間Viterbi algorithm 維特比算法Von Neumann architecture 馮諾伊曼架構(gòu)Wasserstein GAN/WGAN Wasserstein 生成對(duì)抗網(wǎng)絡(luò) Weak learner弱學(xué)習(xí)器Weight權(quán)重Weight sharing 權(quán)共宇Weighted voting加權(quán)投

56、票法Within-class scatter matrix 類內(nèi)散度矩陣Word embedding 詞嵌入Word sense disambiguation 詞義消歧 Zero-data learning 零數(shù)據(jù)學(xué)習(xí) Zero-shot learning 零次學(xué)習(xí) approximations 近似值 arbitrary隨意的 affine仿射的 arbitrary任意的 amino acid氨基酸 amenable經(jīng)得起檢驗(yàn)的concrete具體的事物，實(shí)在的東西 cross validation 交叉驗(yàn)證 correlation相互關(guān)系convention 纟勺立 cluster 一簇 centroids質(zhì)心，形心 converge 收斂 computationally 訃算（機(jī)）的 calculus 計(jì)算derive獲得，取得axiom公理，原則abstract 提取architecture架構(gòu)，體系結(jié)構(gòu)；建造業(yè) abso