+ All Categories
Home > Documents > 자료구조 1차 과제 - WordPress.com · Figure 10-5. Perceptron diagram . w. is the connection...

자료구조 1차 과제 - WordPress.com · Figure 10-5. Perceptron diagram . w. is the connection...

Date post: 06-Aug-2020
Category:
Upload: others
View: 0 times
Download: 0 times
Share this document with a friend
21
Transcript
Page 1: 자료구조 1차 과제 - WordPress.com · Figure 10-5. Perceptron diagram . w. is the connection weight between the ith input neuron and the jth output neu- ron. is the ith input
Page 2: 자료구조 1차 과제 - WordPress.com · Figure 10-5. Perceptron diagram . w. is the connection weight between the ith input neuron and the jth output neu- ron. is the ith input

2

Page 3: 자료구조 1차 과제 - WordPress.com · Figure 10-5. Perceptron diagram . w. is the connection weight between the ith input neuron and the jth output neu- ron. is the ith input

3

Page 4: 자료구조 1차 과제 - WordPress.com · Figure 10-5. Perceptron diagram . w. is the connection weight between the ith input neuron and the jth output neu- ron. is the ith input

4

Walren McCulloch & Walter Pitts proposed a very simple model of biological neurons

Page 5: 자료구조 1차 과제 - WordPress.com · Figure 10-5. Perceptron diagram . w. is the connection weight between the ith input neuron and the jth output neu- ron. is the ith input

5

The Percepton is one of the simplest ANN architectures, invented in 1957 by Frank Rosenblatt.

Page 6: 자료구조 1차 과제 - WordPress.com · Figure 10-5. Perceptron diagram . w. is the connection weight between the ith input neuron and the jth output neu- ron. is the ith input

6

Page 7: 자료구조 1차 과제 - WordPress.com · Figure 10-5. Perceptron diagram . w. is the connection weight between the ith input neuron and the jth output neu- ron. is the ith input

7

Hebb’s rule : “Cells that fire together, wire together.”=> The connection between two neurons is increased whenever

they have the same output.

- The perceptron does not reinforce connections that lead to the wrong output. - It is fed one training instance at a time, and for each instance it makes its predictions.

Page 8: 자료구조 1차 과제 - WordPress.com · Figure 10-5. Perceptron diagram . w. is the connection weight between the ith input neuron and the jth output neu- ron. is the ith input

8

Page 9: 자료구조 1차 과제 - WordPress.com · Figure 10-5. Perceptron diagram . w. is the connection weight between the ith input neuron and the jth output neu- ron. is the ith input

9

Page 10: 자료구조 1차 과제 - WordPress.com · Figure 10-5. Perceptron diagram . w. is the connection weight between the ith input neuron and the jth output neu- ron. is the ith input

10

Page 11: 자료구조 1차 과제 - WordPress.com · Figure 10-5. Perceptron diagram . w. is the connection weight between the ith input neuron and the jth output neu- ron. is the ith input

11

Page 12: 자료구조 1차 과제 - WordPress.com · Figure 10-5. Perceptron diagram . w. is the connection weight between the ith input neuron and the jth output neu- ron. is the ith input

12

Page 13: 자료구조 1차 과제 - WordPress.com · Figure 10-5. Perceptron diagram . w. is the connection weight between the ith input neuron and the jth output neu- ron. is the ith input

13

Page 14: 자료구조 1차 과제 - WordPress.com · Figure 10-5. Perceptron diagram . w. is the connection weight between the ith input neuron and the jth output neu- ron. is the ith input

14

TF.Learn offers a Scikit-Learn-compatible API.

Page 15: 자료구조 1차 과제 - WordPress.com · Figure 10-5. Perceptron diagram . w. is the connection weight between the ith input neuron and the jth output neu- ron. is the ith input

15

MNIST dataset

The first step : construction phase

The second step : execution phase

Page 16: 자료구조 1차 과제 - WordPress.com · Figure 10-5. Perceptron diagram . w. is the connection weight between the ith input neuron and the jth output neu- ron. is the ith input

16

Page 17: 자료구조 1차 과제 - WordPress.com · Figure 10-5. Perceptron diagram . w. is the connection weight between the ith input neuron and the jth output neu- ron. is the ith input

17

Page 18: 자료구조 1차 과제 - WordPress.com · Figure 10-5. Perceptron diagram . w. is the connection weight between the ith input neuron and the jth output neu- ron. is the ith input

18

Page 19: 자료구조 1차 과제 - WordPress.com · Figure 10-5. Perceptron diagram . w. is the connection weight between the ith input neuron and the jth output neu- ron. is the ith input

19

Page 20: 자료구조 1차 과제 - WordPress.com · Figure 10-5. Perceptron diagram . w. is the connection weight between the ith input neuron and the jth output neu- ron. is the ith input

20

Next Step : Fine-Tuning NN Hyperparameters

Random search, Grid search …

Page 21: 자료구조 1차 과제 - WordPress.com · Figure 10-5. Perceptron diagram . w. is the connection weight between the ith input neuron and the jth output neu- ron. is the ith input

21

It has a actually been shown that an MLP with just one hidden layer can model even the most complex functions provided it has enough neurons.

But! It is the fact that deep networks have a much higher parameter efficiency than shallow ones: they can model complex functions using exponentially fewer neurons than shallow nets, making them much faster to train.


Recommended