Scribd
Upload
26 views14 pages

Pertemuan 14 Aplikasi Back Propagation: Matakuliah: H0434/Jaringan Syaraf Tiruan Tahun: 2005 Versi: 1

This document discusses backpropagation and its application in neural networks. Specifically, it demonstrates using a 1-2-1 neural network with backpropagation to approximate a sine wave function. It shows the initial network configuration, forward propagation to get outputs, calculation of error, backpropagation to calculate gradients, and weight updates. It also discusses choosing the appropriate network architecture and shows the network converging on the function and generalizing to new input values.

Uploaded by

danilfikom
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
26 views14 pages

Pertemuan 14 Aplikasi Back Propagation: Matakuliah: H0434/Jaringan Syaraf Tiruan Tahun: 2005 Versi: 1

This document discusses backpropagation and its application in neural networks. Specifically, it demonstrates using a 1-2-1 neural network with backpropagation to approximate a sine wave function. It shows the initial network configuration, forward propagation to get outputs, calculation of error, backpropagation to calculate gradients, and weight updates. It also discusses choosing the appropriate network architecture and shows the network converging on the function and generalizing to new input values.

Uploaded by

danilfikom
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
You are on page 1/ 14

Matakuliah

Tahun
Versi

: H0434/Jaringan Syaraf Tiruan


: 2005
:1

Pertemuan 14
APLIKASI BACK PROPAGATION

Learning Outcomes
Pada akhir pertemuan ini, diharapkan mahasiswa
akan mampu :

Mendemonstrasikan Back Propagation


pada pengenalan pola.

Outline Materi

Aplikasi Back Propagation

Example: Function
Approximation

g p = 1 + sin p
4

+
1-2-1
Network

Network

1-2-1
Network

Initial Conditions
W1 0 = 0.27
0.41

b 1 0 = 0.48
0.13

W2 0 = 0.09 0.17

b 2 0 = 0.48

NetworkResponse
SineWave

-1
-2

-1

Forward Propagation
0

a = p = 1

0.27

0.75

0.48
a = f W a + b = l og sig
= log si g

1 +
0.41
0.13
0.54
1

1 0

0.75
1
1
+
e
a =
= 0.321
1
0.368
0.54

1+ e
2
2
2 1
2
a = f W a + b = purelin ( 0.09 0.17 0.321 + 0.48 ) = 0.446
0.368


2

e = t a = 1 + sin p a = 1 + sin 1 0.446 = 1.261
4
4

Transfer Function Derivatives

d
1
e
1
fn =
= = 1

n
d n 1 + e n
n 2
1 +e
1 + e
1

1 = 1 a 1 a 1

n
1+e

2
d
fn =
n = 1
dn

Backpropagation
2 2
2
2
s = 2 F(n ) t a = 2 f n2 1.261 = 2 1 1.261 = 2.522

2 T 2

s 1 = F(n 1 ) W s =

0
1
s = 1 0.321 0.321

s 1 = 0.218
0

1 a 1 a 1

0.09
2.522
1
1
1 a 2 a 2 0.17

0
0.09
2.522
1 0.368 0.368 0.17

0
0.227 = 0.0495
0.233 0.429
0.0997

Weight Update
= 0.1
T

W 2 1 = W 2 0 s2 a1 = 0.09 0.17 0.1 2.522 0.321 0.368


W2 1 = 0.171 0.0772

b 2 1 = b 2 0 s2 = 0.48 0.1 2.522 = 0.732

T
W1 1 = W1 0 s 1 a 0 = 0.27 0.1 0.0495 1 = 0.265

0.41

0.0997

0.420

b 1 1 = b 1 0 s1 = 0.48 0.1 0.0495 = 0.475


0.13

0.0997

0.140

10

Choice of Architecture
i
g p = 1 + sin p
4

1-3-1 Network
3
2

i=1

-1
-2

-1

3
2

-1
-2

-1

i=4

-1
-2

i=2

-1

-1
-2

i=8

-1

11

Choice of Network
Architecture
6
g p = 1 + sin p
4
3
2

1-2-1

-1
-2

-1

3
2

-1
-2

-1

1-4-1

-1
-2

1-3-1

-1

-1
-2

1-5-1

-1

12

Convergence
g p = 1 + sin p
3

1
3

2
0

2
0

0
-1
-2

-1

-1
-2

1
-1

13

Generalization
{p 1, t 1} { p 2, t2 } {p Q, tQ}

g p = 1 + sin p
4

p = 2 1.6 1.2 1.6 2

1-2-1

1-9-1

-1
-2

-1

-1
-2

-1

14

You might also like