DEEP LEARNING
A movie for everyone
Content
• Motivation

• Machines that learn

• Machine learning history

• Binary classifier

• Linear models and tricks

• Neural Network

• Deep learning
Motivation / Google Trends
Motivation / Google Trends
Jan 2012 - Big Data
Motivation / Google Trends
Oct 2012 - DL
Motivation / Google Trends
Dec 2014 - DL vs NN
Motivation / Google Trends
Ago 2015 - ML vs AI
Motivation / AI landscape
Motivation / AI landscape
Machines that learn
A computer program is said to learn from experience E
with respect to some class of tasks T and performance measure P
if its performance at tasks in T, as measured by P,
improves with experience E.
Tom M. Mitchell
Machines that learn
MACHINE
PROGRAM
INPUT OUTPUT
Machines that learn
MACHINE OUTPUT
INPUT

( FEATURE )
PROGRAM
Experience == Accumulated knowledge ( labelled data )
Machines that learn
MACHINE
OUTPUT

( Prediction )
INPUT

( Any )
PROGRAM
Machine Learning History
Pattern Recognition Machine Learning Deep Learning …
(*)
It is a set of algorithms mainly based on linear models.
Binary Classifier
Binary Classifier
Vegetable
(1,4,3,1,4,3,5,4,6,7,4,5)
Binary Classifier
Animal
(0,1,0,2,3,4,0,6,7,8,1,0)
Binary Classifier
Binary Classifier
Binary Classifier
Binary Classifier
Binary Classifier
Binary Classifier
Binary Classifier
Binary Classifier
Binary Classifier
Binary Classifier
Binary Classifier
Binary Classifier
Binary Classifier
( 0, …. )
( 1, …. )
(1,1,2,3,5,3,2,1 )
Linear Models
XOR problem
Y X1 X2
Blue 1 1
Red 0 1
Blue 0 0
Red
1 0
(0,1)
(0,0)
(1,1)
(1,0)
Linear Models
XOR problem
(0,1)
(0,0)
(1,1)
(1,0)
Linear Models
Linear Models
Y X1 X2 X3
Blue 1 1 0
Red 0 1 1
Blue 0 0 1
Red 1 0 0 Human did classification!
Linear Models
Linear Models
SUPPORT VECTOR MACHINES
Neural Network
Let's mimic what Mother Nature did.
Neural Network
Neural Network
Feed forward is required

to make things simpler.
Neural Network
Neural Network
+1
x1
x2
A
B
C
+1
20
-30
20
20
10
-20
-20
-10
20
x1 x2 A B C
0 0 0 1 1
0 1 0 0 0
1 0 0 0 0
1 1 1 0 1
Neural Network
x1 x2 A B C
0 0 0 1 1
0 1 0 0 0
1 0 0 0 0
1 1 1 0 1
(0,1)
(0,0)
(1,1)
(1,0)
Neural Network
x1 x2 A B C
0 0 0 1 Blue
E
0 1 0 0 Red
1 0 0 0 Red
1 1 1 0 Blue
(0,1)
(0,0)
(1,1)
(1,0)
Deep Learning
Deep Learning
Deep NN, GPUs and 

large amount of data
Deep Learning
Performance
Amount of data
Deep learning
Most ML algorithms
Deep Learning
Big Data
Data Science
Data Mining
AI
Machine Learning
Deep Learning
Deep Learning

Deep learning