鉴于研究生阶段研究的方向,最近正在学习吴恩达的机器学习,本文旨在整理自己学习过程中的英语单词,希望对以后有所帮助。
-----01章 初识机器学习
Application 应用
Computer vision 计算机视觉
Rapid advances in xxx 在xxx方面的飞速发展
Brand 标志
Invent 创造出
Neural network architectures and algorithms 神经网络结构与算法
Cross-fertilization into other areas 在别的领域的交叉成果
Speech recognition 语音识别
Take inspiration from xxx 从xxx得到灵感
Classification 分类
Figure out 识别、弄清楚
Object detection 目标检测
Multiple 许多的
Repaint 重绘
Content 满意的
One of the challenges of computer vision problems is that the input can get really big
Previous 以前的
The input features has dimension xxx 特征向量有xxx维度
特征向量X维度??
Pixel 像素
Megapixel 百万像素
RGB channel RGB颜色模型通道
Total number of weight 所有的权值
Layer 层
Matrix 矩形
A standard fully connected network 标准的全连通网络
Billion parameters 百万参数
Overfit 过拟合
Infeasible 不可接受
Stuck 卡住
Implement 实行
Convolution operation which is one of the fundamental building blocks of convolutional nerual networks
Machine learning practitioners 机器学习从业者
Without being explicitly programmed 没有明确的编程
Remarkable 引入注意的
Slightly more recent definition (更加新的定义 slight轻微的)
Come up with this definition 想到这个定义
Machine learning:A computer program is said to learn from experience E with respect to some task T and some performance measured P , if its performance on T,as measured by P, improves with experience E.
experience E(程序自己与自己下棋的经验)
task T(赢对手)
performance measured性能度量(与新对手下棋赢的概率)
Opponent对手
Content 内容
So,our system’s performance on the task T,on the performance measure P will improve after the exprience E.
所以我们的系统在任务T上的性能,在得到经验E之后回提高性能度量P
On top is 上面是
Filter spam e-mail过滤垃圾邮件
The number of ..... the fraction of ...... xxxx的比例
Machine learning algorithms机器学习算法
01 Supervised learning 监督学习 02unsupervised learning 无监督
Term术语 buzz terms热词
Whereas 而
Reinforcement learning 强化学习 recommend systems推荐系统
(都是机器学习算法,但是最常见的还是监督和无监督学习)
Hammer榔头 screwdriver螺丝刀 saw锯子
Modify 修改
Prediction 预测
Plot策划
Horizontal axis 横轴
Vertical axis竖轴
Quadratic function 二次函数
second-order polynomial 二阶多项式
Refer to指的是
Terminology学术
Regression problem回归问题
Discrete value 离散值
Scalar value标量
Continuous value连续量
Attribute 属性
Possibility可能性
Classification problem 分类问题
Malignant 恶性的
Benign良性的
Cancer type two 第二种肿瘤
Denote表示
Separate out分离
Feature 特点
Infinite 无穷的
Slide幻灯片
Store 存储
Run out of memory耗尽内存
An algorithm called Support Vector Machine 支持向量机算法
Turns out事实证明
Recap 概括
Treat x as y 把x视为y
Real value 实数
Continuous value 连续值
Regression problem(连续值) 、classification problem (离散值0/1)
In supervise learning,in every example in our data set,we are told what is the “correct answer ” that we would have quite liked the algorithms have predicted on that example
(在监督学习中,对于数据集中的每个样本,我们想要算法预测并得出正确答案)
Regression problem,by regression,that means that our goal is to predict continuous valued output
(回归问题,我们的目标是预测一个连续值输出)
classification problem,the goal is to predict a discrete valued output.
(分类问题,目标是预测离散值输出)
Label标记
So-called correct answer 所谓正确的答案
Structure结构
Cluster 簇
Degree程度
Categories类
In advance 提前
A bunch of data一堆数据
Automatically 自动地
We’re not giving the algorithm the right answer for the examples in my data set,this is Unsupervised Learning
(我们没有把例子中数据集中的正确答案,这就是无监督学习)
Social network analysis社交网络分析
Cohesive group of friends同一个圈子的朋友
Unsupervised learning, we have all this customer data but we don’t know in advance what are the market segments.
(无监督学习,我们知道所有用户数据,但我们预先不知道有哪些细分市场)
Organize computer clusters、social network analysis,Market segmentation ,astronomical data analysis
Get rid of摆脱、剔除
Implement实现
Octave programming environment Octave 编译环境
Octave is free open source software 免费的开源软件
Prototype原型
Prototyping tool原型工具
Stands for singular value decomposition 奇异值分解
Stuff 材料
Migrate 迁移
Instructor 导师
Wrap up 总结
------- 02章:单变量相性回归
Liner regression 线性回归
The overall process of x x的整个过程
区分监督学习和无监督学习--看是否有“正确答案”和已知的预测值
Cost function代价函数
M--denote the number of training examples表示训练样本的数量
Lowercase x 小写字母x
Output variables 输出变量
Training set训练集
Hypothesis 假设
Corresponding 相应的
Represent表示
Subscript下标 plus加
Shorthand缩写
Linear regression 线性回归(univariate单变量)
Figure out弄清楚
Straight line直线
Parameters of the model模型参数
Cost function 代价函数(the squared error function平方误差函数)
Mathematical definition数学定义
get back to 回去
Intuition直觉
Recap复习
Form 形成
Optimization objective 优化目标
Visualize可视化
Work with 与...合作
Theta Θ
To minimize J of theta one 减少J(Θ_1)的值
Corresponds to相当于
Simplified definition 简化的定义
Pass through the point(0,0) 过点(0,0)
Concept概念
Hypothesis function假设函数
Θ_1,which controls the slope of the straight line它控制着直线的斜率
Temporary 暂时
Compute计算
One over 2m of my usual cost function 代价函数的1/2m倍
Square 平方
Vertical distance 垂直距离
The predicted value h of x i 预测值h(x^i)
Example样本
Math error计算错误
Flat line 水平线
Negative value 负数
Minus减
By computing the range of values 通过一系列数值的计算
For each value of theta one corresponds to a different hypothesis
每一个Θ_1都对应一个不同的假设函数
Trace out 追踪
Minimize 最小化
Assume 假设
Be familiar with x 对x熟悉
Contour plot 高等线
Contour figures高等图像
illustrator 图像
make sense to 对xx有意义
Problem 问题、课题、难题
Generate生成
Bowl shaped function 碗状函数
3-D surface plot 3-D曲面图
axes 、axis 轴
Vary 改变
Rotate this plot around 旋转这个图形
Ovals、Ellipse 椭圆
The middle of the these concentric ellipse同心椭圆的中心
Intersect 相交
Manually read off the numbers 手工读出数
High dimensional figures with more parameters 具有更多参数、更高维的图形
Gradient descent for minimizing the cost function J 代价函数J最小化的梯度下降法
Arbitrary 随意的
Setup 体系、概述
J(θ0,θ1)是代价函数
For solving this more general problem 为了去解决更一般的问题
For the sake of brevity 简短起见
The sake of 为了
Succinctness 简洁
Notation 符号
Pretend 假装
Wind up 直到、结束
The height of surface 曲面的高度
Pick挑
initialize 初始化
Hill 山
Landscape 景色
Grassy park 青草公园
Spin 360 degrees around and just look all around us旋转360度,看看我们的周围
If I were to take a little baby step in some direction ,and I want to go downhill as quickly as possible,what direction do I take that little baby step in if I want to physically walk down this hill as rapidly as possible?
如果我要在某个方向上走一步,并且我想尽快下山的话,我应该朝什么方向迈步?
Converge、convergence汇合
Property 属性、特点
Local optimum 局部最佳
Intuition 直觉
Subtract 减去
Equation 公式
Detail 细节
Unpack 解压、解释
Assignment 分配、赋值
Assignment operator 赋值运算符
Take the value in b 取b的值
ing