Machine Learning Introduction

1
 Artificial Intelligence

2 2
 What is Artificial Intelligence?

“The capability of a machine to imitate

intelligent human behavior“

3 3
 Early 1950 Optimism began

Artificial
Intelligence
1980’s Machine Learning begins to flourish

Machine
Learning
2010 Deep Learning breakthroughs

Deep
Learning

4
4
5 5
 “

6
6
 “
Machine learning is a field of computer science that gives
computers the ability to learn without being explicitly
programmed.

—Arthur Samuel, 1959

7
 “Classical” decision making

(explicit instructions)
Feature Input Procedure Output
F0 [0.8] if F1 > 0.5 and F2 * F3 < 0.3:
“A”
F1 [0.2] if (F4 – F5) / F6 < 1:
do A
F2 [0.9]
else: “B”
F3 [0.2] if F7 * F0 < 0.3:
F4 [0.0] do B
else: “C”
F5 [0.4]
do C
F6 [0.3] else:
do D “D”
F7 [0.1]

Requires ‘a priori’ knowledge

8
8
 ML decision making

Feature Input Procedure Output

F0 [0.8]
“A”
F1 [0.2]
F2 [0.9] Output = MATRIX * Input
“B”
F3 [0.2] (Linear Regression)

F4 [0.0]
F5 [0.4]
Output = g( M2 * f( M1 * Input) “C”
) knowledge
Requires no (or very little) ‘a priori’
F6 [0.3] (Neural Network with one hidden layer)
“D”
F7 [0.1]

9
9
 Applications of Machine Learning
Data Mining – Web Click
Financial Predictions, Data, Medical Records,
Fraud Detections Diagnosis

Player ranking in Computer vision

online games (barcode, fingerprint and
number plate readers)

Churn Prediction, Voice recognition,

Customer Behaviour Handwriting Recognition
Analysis

10
10
 Source – ANDREW NG

11 11
12
 Machine Learning Cloud Platforms -
Microsoft Azure ML Studio
https://azure.microsoft.com/en-us/services/machine-
learning-studio/

Amazon Machine Learning

https://aws.amazon.com/aml/

SAP Leonardo Machine Learning

https://www.sap.com/india/products/leonardo/machin
e-learning.html

Google ML Platform
https://cloud.google.com/products/machine-learning/

IBM Machine Learning

https://www.ibm.com/analytics/data-science/machine-
learning

13
13
What to learn in machine
Learning?
Programming
and 30%
Tools
Python/R, spark etc.

The Math
behind 70%
Machine
Learning

Probabilistic
Theory, Statistics
and Linear
Algebra

14
14
What to learn in Machine Learning?

Python for Packages for Machine Learning

Machine Learning – Tensorflow, keras etc.

Handling Big Data

Practice, practice & practice

Optimization,
boosting
Techniques

Statistics for Data Science

Math behind
Machine Learning
Algorithms

15 15
 Supervised Unsupervise Reinforcement
Learning d Learning Learning

Learning with a Discovering patterns Learning based on

labeled training set. in unlabeled data. feedback or reward.
Email spam detector Cluster similar Learn to play chess
with training set of documents based on by winning or losing.
already labeled the text content.
emails.

16 16
 Supervised Learning

• We know what we are trying to predict. We use some examples that we (and the model) know the
answer to, to “train” our model. It can then generate predictions to examples we don’t know the
answer to.

• Examples: Predict the price a house will sell at. Identify the gender of someone based on a
photograph.

17 17
 Unsupervised Learning

• We don’t know what we are trying to predict. We are trying to identify some naturally occurring
patterns in the data which may be informative.

• Examples: Try to identify “clusters” of customers based on data we have on them

Y-Values
4

0
0 2 4 6

18 18
 Reinforcement Learning

• Reinforcement learning systems can do multiple things simultaneously -- learn by performing a trial
and error search, learn the model of the environment it is in, and then use that model to plan the
next steps.

• Example: Let's consider a robot whose job is to explore a new building. It has to make sure it has
enough power left to come back to the base station. This robot has to decide if it should make
decisions by considering the trade off between the amount of information collected and the ability
to reach back to base station safely.

19 19
 Types of Problems in Supervised Machine Learning -

Classification Regression
(discrete set of possible outcomes) (possible outcome can by any numerical value with in a
particular continuous range)

20 20
 Types of Problems in Unsupervised Machine Learning -

Clustering Anomaly Detection

(categorization of samples (detecting an anomaly in a general
based on similarity in features) pattern)

21 21
Deep Learning

22 22
 Deep Learning

Deep Learning is part of the machine learning field of

learning representations of data. Exceptional effective
at learning patterns.

23 23
24 24
 Thank you

25