TomMitchell机器学习全部讲义资源-CSDN文库

共13个文件

pdf：13个

需积分: 10 60 浏览量 2018-05-11 13:34:34 上传评论收藏 1.29MB RAR 举报

资源详情

资源评论

资源推荐

收起资源包目录

package

machine Learning.rar （13个子文件）

folder

machine Learning

ch8.pdf 122KB

ch7.pdf 151KB

ch9.pdf 163KB

ch10.pdf 151KB

ch4.pdf 404KB

ch1.pdf 373KB

ch3.pdf 181KB

ch5.pdf 152KB

ch13.pdf 121KB

ch11.pdf 40KB

ch2.pdf 160KB

ch12.pdf 145KB

ch6.pdf 249KB

Articial Neural Networks

[Read Ch. 4]

[Recommended exercises 4.1, 4.2, 4.5, 4.9, 4.11]



Threshold units



Gradient descent



Multilayer networks



Backpropagation



Hidden layer representations



Example: Face Recognition



Advanced topics

74 lecture slides for textbo ok

Machine Learning

, T. Mitchell, McGraw Hill, 1997

Connectionist Mo dels

Consider humans:



Neuron switching time ~

:

001 second



Numb er of neurons ~ 10

10



Connections p er neuron ~ 10

4

?

5



Scene recognition time ~

:

1 second



100 inference steps do esn't seem like enough

!

much parallel computation

Prop erties of articial neural nets (ANN's):



Many neuron-like threshold switching units



Many weighted interconnections among units



Highly parallel, distributed pro cess



Emphasis on tuning weights automatically

75 lecture slides for textbo ok

Machine Learning

, T. Mitchell, McGraw Hill, 1997

When to Consider Neural Networks



Input is high-dimensional discrete or real-valued

(e.g. raw sensor input)



Output is discrete or real valued



Output is a vector of values



Possibly noisy data



Form of target function is unknown



Human readability of result is unimp ortant

Examples:



Sp eech phoneme recognition [Waib el]



Image classication [Kanade, Baluja, Rowley]



Financial prediction

76 lecture slides for textbo ok

Machine Learning

, T. Mitchell, McGraw Hill, 1997

ALVINN drives 70 mph on highways

Sharp

Left

Sharp

Right

4 Hidden

Units

30 Output

Units

30x32 Sensor

Input Retina

Straight

Ahead

77 lecture slides for textbo ok

Machine Learning

, T. Mitchell, McGraw Hill, 1997

Perceptron

w

1

w

2

w

n

w

0

x

1

x

2

x

n

x

0

=1

.

.

.

Σ

AA

Σ w

i

x

i

n

i=0

1 if > 0

-1 otherwise

{

o =

Σ w

i

x

i

n

i=0

o

(

x

1

; : : : ; x

n

) =

8

>

>

<

>

>

:

1 if

w

0

+

w

1

x

1

+

  

+

w

n

x

n

>

0

?

1 otherwise.

Sometimes we'll use simpler vector notation:

o

(

~x

) =

8

>

>

<

>

>

:

1 if

~w



~x >

0

?

1 otherwise.

78 lecture slides for textbo ok

Machine Learning

, T. Mitchell, McGraw Hill, 1997

评论0

内容反馈

sunmoonyzy

粉丝: 0
资源: 2

最新资源

资源上传下载、课程学习等过程中有任何疑问或建议，欢迎提出宝贵意见哦~我们会及时处理！点击此处反馈

feedback-tip