1 of 27

Introduction to Machine Learning

Atul Nag

Adapted from DeepLearning.AI

Adapted from DeepLearning.AI

Adapted from DeepLearning.AI

2 of 27

Machine Learning Overview

What is Machine Learning?

Adapted from DeepLearning.AI

Adapted from DeepLearning.AI

3 of 27

Applications

Web Search – Page Ranking

Upload Pics to Instagram – tag/label friends

OTT similar movies

Voice to text messages

Google Assistant/Siri

Email Spammer flagging

Climate Change

Healthcare

Computer Vision

Computers to learn without explicitly programming

Adapted from DeepLearning.AI

4 of 27

Machine Learning

Applications of machine learning

Adapted from DeepLearning.AI

Adapted from DeepLearning.AI

5 of 27

Adapted from DeepLearning.AI

Adapted from DeepLearning.AI

6 of 27

Applications of Machine Learning

  • State of the Art – Speech Recognition
  • Subfield of Artificial Intelligence
  • Explicit Programming not required
  • Artificial General Intelligence
  • Retail
  • Automation
  • Important to learn the skill set of ML

This Photo by Unknown Author is licensed under CC BY-NC

Adapted from DeepLearning.AI

7 of 27

Machine learning

  • “Field of study that gives computers the ability to learn without being explicitly programmed.”��-- Arthur Samuel (1959)
  • Checkers Playing Program – 10,000 games

Adapted from DeepLearning.AI

8 of 27

Question

If the checkers program had been allowed to play only ten games (instead of tens of thousands) against itself, a much smaller number of games, how would this have affected its performance?

Would have made it better

Would have made it worse

Adapted from DeepLearning.AI

9 of 27

Machine learning algorithms

Supervised Learning

Rapid advancements used most in real world applications

Unsupervised Learning

Recommender Systems

Reinforcement Learning

Adapted from DeepLearning.AI

10 of 27

Machine Learning Overview

Supervised Learning (Part – 1)

Adapted from DeepLearning.AI

Adapted from DeepLearning.AI

11 of 27

Supervised Learning

  • Learns from being given “right answers"

X

    • input

Y

    • Output label

Adapted from DeepLearning.AI

12 of 27

Supervised Learning

Input (X)

Output (Y)

Application

email

Spam? (0/1)

Spam filtering

audio

Text transcripts

Speech recognition

English

Spanish

Machine translation

Ad, user info

Click? (0/1)

Online advertising

Image, radar info

Position of other cars

Self-driving car

Image of phone

Defect? (0/1)

Visual inspection

Adapted from DeepLearning.AI

13 of 27

Regression: Housing price prediction

  • Regression:
    • Predict a number
    • infinitely many possible outputs

​

Adapted from DeepLearning.AI

14 of 27

Classification: Breast cancer detection

Adapted from DeepLearning.AI

15 of 27

Classification: Breast cancer detection

  • Classification – class / category
  • Predict categories – cat / dog, benign / malignant, 0 / 1 / 2 (Can be non-numeric)
  • Small number of possible outputs

Adapted from DeepLearning.AI

16 of 27

Two or more inputs

  • Learning Algorithm sets the boundary

Adapted from DeepLearning.AI

17 of 27

Supervised Learning

Learns from being given “right answers”

  • Regression
    • Predict a number
    • infinitely many possible outputs
  • Classification
    • Predict categories
    • Small number of possible outputs

Adapted from DeepLearning.AI

18 of 27

Quiz

  • Supervised learning is when we give our learning algorithm the right answer y  for each example to learn from.  Which is an example of supervised learning?
  • Calculating the average age of a group of customers.
  • Spam filtering.

Adapted from DeepLearning.AI

19 of 27

Unsupervised Learning

Adapted from DeepLearning.AI

Adapted from DeepLearning.AI

20 of 27

  • Supervised Learning – Learn from data labelled with “right answers”
  • Unsupervised Learning – Find something interesting in unlabelled data

Adapted from DeepLearning.AI

21 of 27

Clustering: Google News

Adapted from DeepLearning.AI

22 of 27

Clustering: DNA Microarray

Adapted from DeepLearning.AI

23 of 27

Clustering: Grouping customers

Adapted from DeepLearning.AI

24 of 27

Unsupervised Learning

  • Data only comes with inputs x, but not output labels y.
  • Algorithm has to find structure in the data.

Clustering

Group similar data points together.

Dimensionality reduction

Compress data using fewer numbers

Anomaly detection

Find unusual data points

Adapted from DeepLearning.AI

25 of 27

Of the following examples, which would you address using an unsupervised learning algorithm?

  • Given email labelled as spam/not spam, learn a spam filter
  • Given a set of news articles found on the web, group them into sets of articles about the same story
  • Given a database of customer data, automatically discover market segments and group customers into different market segments
  • Given a dataset of patients diagnosed as either having diabetes or not, learn to classify new patients as having diabetes or not

Adapted from DeepLearning.AI

26 of 27

Quiz - 1

  • Which are the two common types of supervised learning? (Choose two)
  • Classification
  • Regression
  • Clustering

Adapted from DeepLearning.AI

27 of 27

Quiz - 2

  • Which of these is a type of unsupervised learning?
  • Classification
  • Regression
  • Clustering

Adapted from DeepLearning.AI