activation 激活值 

activation function 激活函数 

additive noise 加性噪声 

autoencoder 自编码器 

Autoencoders 自编码算法 

average firing rate 平均激活率 

average sum-of-squares error 均方差 

backpropagation 后向传播 

basis 基 

basis feature vectors 特征基向量 

batch gradient ascent 批量梯度上升法 

Bayesian regularization method 贝叶斯规则化方法 

Bernoulli random variable 伯努利随机变量 

bias term 偏置项 

binary classfication 二元分类 

class labels 类型标记 

concatenation 级联 

conjugate gradient 共轭梯度 

contiguous groups 联通区域 

convex optimization software 凸优化软件 

convolution 卷积 

cost function 代价函数 

covariance matrix 协方差矩阵 

DC component 直流分量 

decorrelation 去相关 

degeneracy 退化 

demensionality reduction 降维 

derivative 导函数 

diagonal 对角线 

diffusion of gradients 梯度的弥散 

eigenvalue 特征值 

eigenvector 特征向量 

error term 残差 

feature matrix 特征矩阵 

feature standardization 特征标准化 

feedforward architectures 前馈结构算法 

feedforward neural network 前馈神经网络 

feedforward pass 前馈传导 

fine-tuned 微调 

first-order feature 一阶特征 

forward pass 前向传导 

forward propagation 前向传播 

Gaussian prior 高斯先验概率 

generative model 生成模型 

gradient descent 梯度下降 

Greedy layer-wise training 逐层贪婪训练方法 

grouping matrix 分组矩阵 

Hadamard product 阿达马乘积 

Hessian matrix Hessian 矩阵 

hidden layer 隐含层 

hidden units 隐藏神经元 

Hierarchical grouping 层次型分组 

higher-order features 更高阶特征 

highly non-convex optimization problem 高度非凸的优化问题 

histogram 直方图 

hyperbolic tangent 双曲正切函数 

hypothesis 估值，假设 

identity activation function 恒等激励函数 

IID 独立同分布 

illumination 照明 

inactive 抑制 

independent component analysis 独立成份分析 

input domains 输入域 

input layer 输入层 

intensity 亮度/灰度 

intercept term 截距 

KL divergence 相对熵 

KL divergence KL分散度 

k-Means K-均值 

learning rate 学习速率 

least squares 最小二乘法 

linear correspondence 线性响应 

linear superposition 线性叠加 

line-search algorithm 线搜索算法 

local mean subtraction 局部均值消减 

local optima 局部最优解 

logistic regression 逻辑回归 

loss function 损失函数 

low-pass filtering 低通滤波 

magnitude 幅值 

MAP 极大后验估计 

maximum likelihood estimation 极大似然估计 

mean 平均值 

MFCC Mel 倒频系数 

multi-class classification 多元分类 

neural networks 神经网络 

neuron 神经元 

Newton’s method 牛顿法 

non-convex function 非凸函数 

non-linear feature 非线性特征 

norm 范式 

norm bounded 有界范数 

norm constrained 范数约束 

normalization 归一化 

numerical roundoff errors 数值舍入误差 

numerically checking 数值检验 

numerically reliable 数值计算上稳定 

object detection 物体检测 

objective function 目标函数 

off-by-one error 缺位错误 

orthogonalization 正交化 

output layer 输出层 

overall cost function 总体代价函数 

over-complete basis 超完备基 

over-fitting 过拟合 

parts of objects 目标的部件 

part-whole decompostion 部分-整体分解 

PCA 主元分析 

penalty term 惩罚因子 

per-example mean subtraction 逐样本均值消减 

pooling 池化 

pretrain 预训练 

principal components analysis 主成份分析 

quadratic constraints 二次约束 

RBMs 受限Boltzman机 

reconstruction based models 基于重构的模型 

reconstruction cost 重建代价 

reconstruction term 重构项 

redundant 冗余 

reflection matrix 反射矩阵 

regularization 正则化 

regularization term 正则化项 

rescaling 缩放 

robust 鲁棒性 

run 行程 

second-order feature 二阶特征 

sigmoid activation function S型激励函数 

significant digits 有效数字 

singular value 奇异值 

singular vector 奇异向量 

smoothed L1 penalty 平滑的L1范数惩罚 

Smoothed topographic L1 sparsity penalty 平滑地形L1稀疏惩罚函数 

smoothing 平滑 

Softmax Regresson Softmax回归 

sorted in decreasing order 降序排列 

source features 源特征 

sparse autoencoder 消减归一化 

Sparsity 稀疏性 

sparsity parameter 稀疏性参数 

sparsity penalty 稀疏惩罚 

square function 平方函数 

squared-error 方差 

stationary 平稳性（不变性） 

stationary stochastic process 平稳随机过程 

step-size 步长值 

supervised learning 监督学习 

symmetric positive semi-definite matrix 对称半正定矩阵 

symmetry breaking 对称失效 

tanh function 双曲正切函数 

the average activation 平均活跃度 

the derivative checking method 梯度验证方法 

the empirical distribution 经验分布函数 

the energy function 能量函数 

the Lagrange dual 拉格朗日对偶函数 

the log likelihood 对数似然函数 

the pixel intensity value 像素灰度值 

the rate of convergence 收敛速度 

topographic cost term 拓扑代价项 

topographic ordered 拓扑秩序 

transformation 变换 

translation invariant 平移不变性 

trivial answer 平凡解 

under-complete basis 不完备基 

unrolling 组合扩展 

unsupervised learning 无监督学习 

variance 方差 

vecotrized implementation 向量化实现 

vectorization 矢量化 

visual cortex 视觉皮层 

weight decay 权重衰减 

weighted average 加权平均值 

whitening 白化 

zero-mean 均值为零

Letter A

Accumulated error backpropagation 累积误差逆传播 

Activation Function 激活函数 

Adaptive Resonance Theory/ART 自适应谐振理论 

Addictive model 加性学习 

Adversarial Networks 对抗网络 

Affine Layer 仿射层 

Affinity matrix 亲和矩阵 

Agent 代理 / 智能体 

Algorithm 算法 

Alpha-beta pruning α-β剪枝 

Anomaly detection 异常检测 

Approximation 近似 

Area Under ROC Curve／AUC Roc 曲线下面积 

Artificial General Intelligence/AGI 通用人工智能 

Artificial Intelligence/AI 人工智能 

Association analysis 关联分析 

Attention mechanism 注意力机制 

Attribute conditional independence assumption 属性条件独立性假设 

Attribute space 属性空间 

Attribute value 属性值 

Autoencoder 自编码器 

Automatic speech recognition 自动语音识别 

Automatic summarization 自动摘要 

Average gradient 平均梯度 

Average-Pooling 平均池化

Letter B

Backpropagation Through Time 通过时间的反向传播 

Backpropagation/BP 反向传播 

Base learner 基学习器 

Base learning algorithm 基学习算法 

Batch Normalization/BN 批量归一化 

Bayes decision rule 贝叶斯判定准则 

Bayes Model Averaging／BMA 贝叶斯模型平均 

Bayes optimal classifier 贝叶斯最优分类器 

Bayesian decision theory 贝叶斯决策论 

Bayesian network 贝叶斯网络 

Between-class scatter matrix 类间散度矩阵 

Bias 偏置 / 偏差 

Bias-variance decomposition 偏差-方差分解 

Bias-Variance Dilemma 偏差 – 方差困境 

Bi-directional Long-Short Term Memory/Bi-LSTM 双向长短期记忆 

Binary classification 二分类 

Binomial test 二项检验 

Bi-partition 二分法 

Boltzmann machine 玻尔兹曼机 

Bootstrap sampling 自助采样法／可重复采样／有放回采样 

Bootstrapping 自助法 

Break-Event Point／BEP 平衡点

Letter C

Calibration 校准 

Cascade-Correlation 级联相关 

Categorical attribute 离散属性 

Class-conditional probability 类条件概率 

Classification and regression tree/CART 分类与回归树 

Classifier 分类器 

Class-imbalance 类别不平衡 

Closed -form 闭式 

Cluster 簇/类/集群 

Cluster analysis 聚类分析 

Clustering 聚类 

Clustering ensemble 聚类集成 

Co-adapting 共适应 

Coding matrix 编码矩阵 

COLT 国际学习理论会议 

Committee-based learning 基于委员会的学习 

Competitive learning 竞争型学习 

Component learner 组件学习器 

Comprehensibility 可解释性 

Computation Cost 计算成本 

Computational Linguistics 计算语言学 

Computer vision 计算机视觉 

Concept drift 概念漂移 

Concept Learning System /CLS 概念学习系统 

Conditional entropy 条件熵 

Conditional mutual information 条件互信息 

Conditional Probability Table／CPT 条件概率表 

Conditional random field/CRF 条件随机场 

Conditional risk 条件风险 

Confidence 置信度 

Confusion matrix 混淆矩阵 

Connection weight 连接权 

Connectionism 连结主义 

Consistency 一致性／相合性 

Contingency table 列联表 

Continuous attribute 连续属性 

Convergence 收敛 

Conversational agent 会话智能体 

Convex quadratic programming 凸二次规划 

Convexity 凸性 

Convolutional neural network/CNN 卷积神经网络 

Co-occurrence 同现 

Correlation coefficient 相关系数 

Cosine similarity 余弦相似度 

Cost curve 成本曲线 

Cost Function 成本函数 

Cost matrix 成本矩阵 

Cost-sensitive 成本敏感 

Cross entropy 交叉熵 

Cross validation 交叉验证 

Crowdsourcing 众包 

Curse of dimensionality 维数灾难 

Cut point 截断点 

Cutting plane algorithm 割平面法

Letter D

Data mining 数据挖掘 

Data set 数据集 

Decision Boundary 决策边界 

Decision stump 决策树桩 

Decision tree 决策树／判定树 

Deduction 演绎 

Deep Belief Network 深度信念网络 

Deep Convolutional Generative Adversarial Network/DCGAN 深度卷积生成对抗网络 

Deep learning 深度学习 

Deep neural network/DNN 深度神经网络 

Deep Q-Learning 深度 Q 学习 

Deep Q-Network 深度 Q 网络 

Density estimation 密度估计 

Density-based clustering 密度聚类 

Differentiable neural computer 可微分神经计算机 

Dimensionality reduction algorithm 降维算法 

Directed edge 有向边 

Disagreement measure 不合度量 

Discriminative model 判别模型 

Discriminator 判别器 

Distance measure 距离度量 

Distance metric learning 距离度量学习 

Distribution 分布 

Divergence 散度 

Diversity measure 多样性度量／差异性度量 

Domain adaption 领域自适应 

Downsampling 下采样 

D-separation （Directed separation） 有向分离 

Dual problem 对偶问题 

Dummy node 哑结点 

Dynamic Fusion 动态融合 

Dynamic programming 动态规划

Letter E

Eigenvalue decomposition 特征值分解 

Embedding 嵌入 

Emotional analysis 情绪分析 

Empirical conditional entropy 经验条件熵 

Empirical entropy 经验熵 

Empirical error 经验误差 

Empirical risk 经验风险 

End-to-End 端到端 

Energy-based model 基于能量的模型 

Ensemble learning 集成学习 

Ensemble pruning 集成修剪 

Error Correcting Output Codes／ECOC 纠错输出码 

Error rate 错误率 

Error-ambiguity decomposition 误差-分歧分解 

Euclidean distance 欧氏距离 

Evolutionary computation 演化计算 

Expectation-Maximization 期望最大化 

Expected loss 期望损失 

Exploding Gradient Problem 梯度爆炸问题 

Exponential loss function 指数损失函数 

Extreme Learning Machine/ELM 超限学习机

Letter F

Factorization 因子分解 

False negative 假负类 

False positive 假正类 

False Positive Rate/FPR 假正例率 

Feature engineering 特征工程 

Feature selection 特征选择 

Feature vector 特征向量 

Featured Learning 特征学习 

Feedforward Neural Networks/FNN 前馈神经网络 

Fine-tuning 微调 

Flipping output 翻转法 

Fluctuation 震荡 

Forward stagewise algorithm 前向分步算法 

Frequentist 频率主义学派 

Full-rank matrix 满秩矩阵 

Functional neuron 功能神经元

Letter G

Gain ratio 增益率 

Game theory 博弈论 

Gaussian kernel function 高斯核函数 

Gaussian Mixture Model 高斯混合模型 

General Problem Solving 通用问题求解 

Generalization 泛化 

Generalization error 泛化误差 

Generalization error bound 泛化误差上界 

Generalized Lagrange function 广义拉格朗日函数 

Generalized linear model 广义线性模型 

Generalized Rayleigh quotient 广义瑞利商 

Generative Adversarial Networks/GAN 生成对抗网络 

Generative Model 生成模型 

Generator 生成器 

Genetic Algorithm/GA 遗传算法 

Gibbs sampling 吉布斯采样 

Gini index 基尼指数 

Global minimum 全局最小 

Global Optimization 全局优化 

Gradient boosting 梯度提升 

Gradient Descent 梯度下降 

Graph theory 图论 

Ground-truth 真相／真实

Letter H

Hard margin 硬间隔 

Hard voting 硬投票 

Harmonic mean 调和平均 

Hesse matrix 海塞矩阵 

Hidden dynamic model 隐动态模型 

Hidden layer 隐藏层 

Hidden Markov Model/HMM 隐马尔可夫模型 

Hierarchical clustering 层次聚类 

Hilbert space 希尔伯特空间 

Hinge loss function 合页损失函数 

Hold-out 留出法 

Homogeneous 同质 

Hybrid computing 混合计算 

Hyperparameter 超参数 

Hypothesis 假设 

Hypothesis test 假设验证

Letter I

ICML 国际机器学习会议 

Improved iterative scaling/IIS 改进的迭代尺度法 

Incremental learning 增量学习 

Independent and identically distributed/i.i.d. 独立同分布 

Independent Component Analysis/ICA 独立成分分析 

Indicator function 指示函数 

Individual learner 个体学习器 

Induction 归纳 

Inductive bias 归纳偏好 

Inductive learning 归纳学习 

Inductive Logic Programming／ILP 归纳逻辑程序设计 

Information entropy 信息熵 

Information gain 信息增益 

Input layer 输入层 

Insensitive loss 不敏感损失 

Inter-cluster similarity 簇间相似度 

International Conference for Machine Learning/ICML 国际机器学习大会 

Intra-cluster similarity 簇内相似度 

Intrinsic value 固有值 

Isometric Mapping/Isomap 等度量映射 

Isotonic regression 等分回归 

Iterative Dichotomiser 迭代二分器

Letter K

Kernel method 核方法 

Kernel trick 核技巧 

Kernelized Linear Discriminant Analysis／KLDA 核线性判别分析 

K-fold cross validation k 折交叉验证／k 倍交叉验证 

K-Means Clustering K – 均值聚类 

K-Nearest Neighbours Algorithm/KNN K近邻算法 

Knowledge base 知识库 

Knowledge Representation 知识表征

Letter L

Label space 标记空间 

Lagrange duality 拉格朗日对偶性 

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 隐狄利克雷分布 

Latent semantic analysis 潜在语义分析 

Latent variable 隐变量 

Lazy learning 懒惰学习 

Learner 学习器 

Learning by analogy 类比学习 

Learning rate 学习率 

Learning Vector Quantization/LVQ 学习向量量化 

Least squares regression tree 最小二乘回归树 

Leave-One-Out/LOO 留一法 

linear chain conditional random field 线性链条件随机场 

Linear Discriminant Analysis／LDA 线性判别分析 

Linear model 线性模型 

Linear Regression 线性回归 

Link function 联系函数 

Local Markov property 局部马尔可夫性 

Local minimum 局部最小 

Log likelihood 对数似然 

Log odds／logit 对数几率 

Logistic Regression Logistic 回归 

Log-likelihood 对数似然 

Log-linear regression 对数线性回归 

Long-Short Term Memory/LSTM 长短期记忆 

Loss function 损失函数

Letter M

Machine translation/MT 机器翻译 

Macron-P 宏查准率 

Macron-R 宏查全率 

Majority voting 绝对多数投票法 

Manifold assumption 流形假设 

Manifold learning 流形学习 

Margin theory 间隔理论 

Marginal distribution 边际分布 

Marginal independence 边际独立性 

Marginalization 边际化 

Markov Chain Monte Carlo/MCMC 马尔可夫链蒙特卡罗方法 

Markov Random Field 马尔可夫随机场 

Maximal clique 最大团 

Maximum Likelihood Estimation/MLE 极大似然估计／极大似然法 

Maximum margin 最大间隔 

Maximum weighted spanning tree 最大带权生成树 

Max-Pooling 最大池化 

Mean squared error 均方误差 

Meta-learner 元学习器 

Metric learning 度量学习 

Micro-P 微查准率 

Micro-R 微查全率 

Minimal Description Length/MDL 最小描述长度 

Minimax game 极小极大博弈 

Misclassification cost 误分类成本 

Mixture of experts 混合专家 

Momentum 动量 

Moral graph 道德图／端正图 

Multi-class classification 多分类 

Multi-document summarization 多文档摘要 

Multi-layer feedforward neural networks 多层前馈神经网络 

Multilayer Perceptron/MLP 多层感知器 

Multimodal learning 多模态学习 

Multiple Dimensional Scaling 多维缩放 

Multiple linear regression 多元线性回归 

Multi-response Linear Regression ／MLR 多响应线性回归 

Mutual information 互信息

Letter N

Naive bayes 朴素贝叶斯 

Naive Bayes Classifier 朴素贝叶斯分类器 

Named entity recognition 命名实体识别 

Nash equilibrium 纳什均衡 

Natural language generation/NLG 自然语言生成 

Natural language processing 自然语言处理 

Negative class 负类 

Negative correlation 负相关法 

Negative Log Likelihood 负对数似然 

Neighbourhood Component Analysis/NCA 近邻成分分析 

Neural Machine Translation 神经机器翻译 

Neural Turing Machine 神经图灵机 

Newton method 牛顿法 

NIPS 国际神经信息处理系统会议 

No Free Lunch Theorem／NFL 没有免费的午餐定理 

Noise-contrastive estimation 噪音对比估计 

Nominal attribute 列名属性 

Non-convex optimization 非凸优化 

Nonlinear model 非线性模型 

Non-metric distance 非度量距离 

Non-negative matrix factorization 非负矩阵分解 

Non-ordinal attribute 无序属性 

Non-Saturating Game 非饱和博弈 

Norm 范数 

Normalization 归一化 

Nuclear norm 核范数 

Numerical attribute 数值属性

Letter O

Objective function 目标函数 

Oblique decision tree 斜决策树 

Occam’s razor 奥卡姆剃刀 

Odds 几率 

Off-Policy 离策略 

One shot learning 一次性学习 

One-Dependent Estimator／ODE 独依赖估计 

On-Policy 在策略 

Ordinal attribute 有序属性 

Out-of-bag estimate 包外估计 

Output layer 输出层 

Output smearing 输出调制法 

Overfitting 过拟合／过配 

Oversampling 过采样

Letter P

Paired t-test 成对 t 检验 

Pairwise 成对型 

Pairwise Markov property 成对马尔可夫性 

Parameter 参数 

Parameter estimation 参数估计 

Parameter tuning 调参 

Parse tree 解析树 

Particle Swarm Optimization/PSO 粒子群优化算法 

Part-of-speech tagging 词性标注 

Perceptron 感知机 

Performance measure 性能度量 

Plug and Play Generative Network 即插即用生成网络 

Plurality voting 相对多数投票法 

Polarity detection 极性检测 

Polynomial kernel function 多项式核函数 

Pooling 池化 

Positive class 正类 

Positive definite matrix 正定矩阵 

Post-hoc test 后续检验 

Post-pruning 后剪枝 

potential function 势函数 

Precision 查准率／准确率 

Prepruning 预剪枝 

Principal component analysis/PCA 主成分分析 

Principle of multiple explanations 多释原则 

Prior 先验 

Probability Graphical Model 概率图模型 

Proximal Gradient Descent/PGD 近端梯度下降 

Pruning 剪枝 

Pseudo-label 伪标记

Letter Q

Quantized Neural Network 量子化神经网络 

Quantum computer 量子计算机 

Quantum Computing 量子计算 

Quasi Newton method 拟牛顿法

Letter R

Radial Basis Function／RBF 径向基函数 

Random Forest Algorithm 随机森林算法 

Random walk 随机漫步 

Recall 查全率／召回率 

Receiver Operating Characteristic/ROC 受试者工作特征 

Rectified Linear Unit/ReLU 线性修正单元 

Recurrent Neural Network 循环神经网络 

Recursive neural network 递归神经网络 

Reference model 参考模型 

Regression 回归 

Regularization 正则化 

Reinforcement learning/RL 强化学习 

Representation learning 表征学习 

Representer theorem 表示定理 

reproducing kernel Hilbert space/RKHS 再生核希尔伯特空间 

Re-sampling 重采样法 

Rescaling 再缩放 

Residual Mapping 残差映射 

Residual Network 残差网络 

Restricted Boltzmann Machine/RBM 受限玻尔兹曼机 

Restricted Isometry Property/RIP 限定等距性 

Re-weighting 重赋权法 

Robustness 稳健性/鲁棒性 

Root node 根结点 

Rule Engine 规则引擎 

Rule learning 规则学习

Letter S

Saddle point 鞍点 

Sample space 样本空间 

Sampling 采样 

Score function 评分函数 

Self-Driving 自动驾驶 

Self-Organizing Map／SOM 自组织映射 

Semi-naive Bayes classifiers 半朴素贝叶斯分类器 

Semi-Supervised Learning 半监督学习 

semi-Supervised Support Vector Machine 半监督支持向量机 

Sentiment analysis 情感分析 

Separating hyperplane 分离超平面 

Sigmoid function Sigmoid 函数 

Similarity measure 相似度度量 

Simulated annealing 模拟退火 

Simultaneous localization and mapping 同步定位与地图构建 

Singular Value Decomposition 奇异值分解 

Slack variables 松弛变量 

Smoothing 平滑 

Soft margin 软间隔 

Soft margin maximization 软间隔最大化 

Soft voting 软投票 

Sparse representation 稀疏表征 

Sparsity 稀疏性 

Specialization 特化 

Spectral Clustering 谱聚类 

Speech Recognition 语音识别 

Splitting variable 切分变量 

Squashing function 挤压函数 

Stability-plasticity dilemma 可塑性-稳定性困境 

Statistical learning 统计学习 

Status feature function 状态特征函 

Stochastic gradient descent 随机梯度下降 

Stratified sampling 分层采样 

Structural risk 结构风险 

Structural risk minimization/SRM 结构风险最小化 

Subspace 子空间 

Supervised learning 监督学习／有导师学习 

support vector expansion 支持向量展式 

Support Vector Machine/SVM 支持向量机 

Surrogat loss 替代损失 

Surrogate function 替代函数 

Symbolic learning 符号学习 

Symbolism 符号主义 

Synset 同义词集

Letter T

T-Distribution Stochastic Neighbour Embedding/t-SNE T – 分布随机近邻嵌入 

Tensor 张量 

Tensor Processing Units/TPU 张量处理单元 

The least square method 最小二乘法 

Threshold 阈值 

Threshold logic unit 阈值逻辑单元 

Threshold-moving 阈值移动 

Time Step 时间步骤 

Tokenization 标记化 

Training error 训练误差 

Training instance 训练示例／训练例 

Transductive learning 直推学习 

Transfer learning 迁移学习 

Treebank 树库 

Tria-by-error 试错法 

True negative 真负类 

True positive 真正类 

True Positive Rate/TPR 真正例率 

Turing Machine 图灵机 

Twice-learning 二次学习

Letter U

Underfitting 欠拟合／欠配 

Undersampling 欠采样 

Understandability 可理解性 

Unequal cost 非均等代价 

Unit-step function 单位阶跃函数 

Univariate decision tree 单变量决策树 

Unsupervised learning 无监督学习／无导师学习 

Unsupervised layer-wise training 无监督逐层训练 

Upsampling 上采样

Letter V

Vanishing Gradient Problem 梯度消失问题 

Variational inference 变分推断 

VC Theory VC维理论 

Version space 版本空间 

Viterbi algorithm 维特比算法 

Von Neumann architecture 冯 · 诺伊曼架构

Letter W

Wasserstein GAN/WGAN Wasserstein生成对抗网络 

Weak learner 弱学习器 

Weight 权重 

Weight sharing 权共享 

Weighted voting 加权投票法 

Within-class scatter matrix 类内散度矩阵 

Word embedding 词嵌入 

Word sense disambiguation 词义消歧

Letter Z

Zero-data learning 零数据学习 

Zero-shot learning 零次学习

A

approximations近似值 

arbitrary随意的 

affine仿射的 

arbitrary任意的 

amino acid氨基酸 

amenable经得起检验的 

axiom公理，原则 

abstract提取 

architecture架构，体系结构；建造业 

absolute绝对的 

arsenal军火库 

assignment分配 

algebra线性代数 

asymptotically无症状的 

appropriate恰当的

B

bias偏差 

brevity简短，简洁；短暂 

broader广泛 

briefly简短的 

batch批量

C

convergence 收敛，集中到一点 

convex凸的 

contours轮廓 

constraint约束 

constant常理 

commercial商务的 

complementarity补充 

coordinate ascent同等级上升 

clipping剪下物；剪报；修剪 

component分量；部件 

continuous连续的 

covariance协方差 

canonical正规的，正则的 

concave非凸的 

corresponds相符合；相当；通信 

corollary推论 

concrete具体的事物，实在的东西 

cross validation交叉验证 

correlation相互关系 

convention约定 

cluster一簇 

centroids 质心，形心 

converge收敛 

computationally计算(机)的 

calculus计算

D

derive获得，取得 

dual二元的 

duality二元性；二象性；对偶性 

derivation求导；得到；起源 

denote预示，表示，是…的标志；意味着，[逻]指称 

divergence 散度；发散性 

dimension尺度，规格；维数 

dot小圆点 

distortion变形 

density概率密度函数 

discrete离散的 

discriminative有识别能力的 

diagonal对角 

dispersion分散，散开 

determinant决定因素 

disjoint不相交的

E

encounter遇到 

ellipses椭圆 

equality等式 

extra额外的 

empirical经验；观察 

ennmerate例举，计数 

exceed超过，越出 

expectation期望 

efficient生效的 

endow赋予 

explicitly清楚的 

exponential family指数家族 

equivalently等价的

F

feasible可行的 

forary初次尝试 

finite有限的，限定的 

forgo摒弃，放弃 

fliter过滤 

frequentist最常发生的 

forward search前向式搜索 

formalize使定形

G

generalized归纳的 

generalization概括，归纳；普遍化；判断（根据不足） 

guarantee保证；抵押品 

generate形成，产生 

geometric margins几何边界 

gap裂口 

generative生产的；有生产力的

H

heuristic启发式的；启发法；启发程序 

hone怀恋；磨 

hyperplane超平面

L

initial最初的 

implement执行 

intuitive凭直觉获知的 

incremental增加的 

intercept截距 

intuitious直觉 

instantiation例子 

indicator指示物，指示器 

interative重复的，迭代的 

integral积分 

identical相等的；完全相同的 

indicate表示，指出 

invariance不变性，恒定性 

impose把…强加于 

intermediate中间的 

interpretation解释，翻译

J

joint distribution联合概率

L

lieu替代 

logarithmic对数的，用对数表示的 

latent潜在的 

Leave-one-out cross validation留一法交叉验证

M

magnitude巨大 

mapping绘图，制图；映射 

matrix矩阵 

mutual相互的，共同的 

monotonically单调的 

minor较小的，次要的 

multinomial多项的 

multi-class classification二分类问题

N

nasty讨厌的 

notation标志，注释 

naïve朴素的

O

obtain得到 

oscillate摆动 

optimization problem最优化问题 

objective function目标函数 

optimal最理想的 

orthogonal(矢量，矩阵等)正交的 

orientation方向 

ordinary普通的 

occasionally偶然的

P

partial derivative偏导数 

property性质 

proportional成比例的 

primal原始的，最初的 

permit允许 

pseudocode伪代码 

permissible可允许的 

polynomial多项式 

preliminary预备 

precision精度 

perturbation 不安，扰乱 

poist假定，设想 

positive semi-definite半正定的 

parentheses圆括号 

posterior probability后验概率 

plementarity补充 

pictorially图像的 

parameterize确定…的参数 

poisson distribution柏松分布 

pertinent相关的

Q

quadratic二次的 

quantity量，数量；分量 

query疑问的

R

regularization使系统化；调整 

reoptimize重新优化 

restrict限制；限定；约束 

reminiscent回忆往事的；提醒的；使人联想…的（of） 

remark注意 

random variable随机变量 

respect考虑 

respectively各自的；分别的 

redundant过多的；冗余的

S

susceptible敏感的 

stochastic可能的；随机的 

symmetric对称的 

sophisticated复杂的 

spurious假的；伪造的 

subtract减去；减法器 

simultaneously同时发生地；同步地 

suffice满足 

scarce稀有的，难得的 

split分解，分离 

subset子集 

statistic统计量 

successive iteratious连续的迭代 

scale标度 

sort of有几分的 

squares平方

T

trajectory轨迹 

temporarily暂时的 

terminology专用名词 

tolerance容忍；公差 

thumb翻阅 

threshold阈，临界 

theorem定理 

tangent正弦

U

unit-length vector单位向量

V

valid有效的，正确的 

variance方差 

variable变量；变元 

vocabulary词汇 

valued经估价的；宝贵的

W

wrapper包装