W3B & GDG

MACHINE
LEARNING
Balancing Innovation and Responsibility
W3B: ML Workshop

WHAT IS MACHINE LEARNING?

Machine Learning (ML) is a branch of

artificial intelligence (AI) that enables
computers to learn from data and
improve their performance on tasks
without being explicitly programmed.
Instead of writing code with fixed rules,
you feed the machine data, and it
identifies patterns to make predictions
or decisions.
 W3B: ML Workshop

TRADITIONAL PROGRAMMING

You will write explicit rules to solve a problem.

Manually written by developers.
Hard to adapt to new data.
Defined Problems with clear rules.
W3B: ML Workshop

WHAT DOES ‘LEARNING’ MEAN IN ML?

The model improves its performance

by finding patterns in data over time.
It generalizes from past examples to
predict unseen data.
The model adjusts its internal
parameters (like weights in neural
networks) to minimize errors and
improve accuracy.
 W3B: ML Workshop

Imagine teaching a child to recognize

apples:

Show them pictures of apples and non-

apples.
They start identifying features (round
shape, red/green color).
Over time, they improve at recognizing
apples without explicit instructions for
every type of apple.
 W3B: ML Workshop

ORIGIN OF MACHINE LEARNING

A long time ago, computers could only do exactly what people told them to do.
Imagine if you had to give your computer every little instruction, like telling it to pick
up a toy or draw a picture. It would be very slow and hard to do everything!

Then, some smart people had a great idea: "What if we teach computers to learn from
experience, just like we do?" So, they started showing computers lots of examples
instead of telling them exactly what to do.

For example, they might show the computer pictures of cats and dogs over and over.
The computer doesn’t know what a cat or dog is at first, but after seeing many
pictures, it starts to figure out the difference on its own. This was the beginning of
machine learning!
 W3B: ML Workshop

ORIGIN OF MACHINE LEARNING

Back in the 1950s, a smart man named Alan
Turing asked, “Can a machine think like a
person?” This made people start thinking
about teaching machines to be smart, and
soon, they started teaching computers to
recognize things and make decisions. By the
time the 2000s came around, machines
could learn a lot faster, and that’s when we
started seeing things like smart phones,
online recommendations, and even self-
driving cars!
 W3B: ML Workshop

ORIGIN OF MACHINE LEARNING

1950s - Early Al models focused on rule-based approaches.

1959 - Arthur Samuel coined the term "machine learning" while developing a system that could
learn to play checkers.

1980s-2000s - Development of neural networks and increased computational power accelerated

progress.

2010s to Present - Massive data availability, GPUs, and deep learning revolutionized ML, enabling
breakthroughs in fields like NLP, computer vision, and robotics.
 W3B: ML Workshop

COMMONLY USED TERMS IN MACHINE LEARNING

There are many commonly used terms like:
1. Data: Data is just information. For machine learning, it’s like the "fuel"
that helps the computer learn. This could be pictures, numbers, words,
or anything the computer uses to learn.
2. Model: A model is like a brain for the computer. It takes the data,
learns from it, and tries to make decisions or predictions.
3. Training: Training is like practice. The computer looks at lots of
examples (data) and adjusts itself to get better at making predictions.
4. Algorithm: An algorithm is like a set of instructions or a recipe. It tells
the computer how to learn from the data.
 W3B: ML Workshop

COMMONLY USED TERMS IN MACHINE LEARNING

Feature: A feature is a piece of information the computer uses to
understand something. It's like a clue that helps the model make a
decision.
Label: A label is the answer or category that we want the computer to
predict. It tells the computer what the data should be.
Training Set and Test Set: The training set is the data the computer uses
to learn, while the test set is the data used to check how well the
computer has learned.
Overfitting: Overfitting happens when the computer learns the details of
the training data too well and becomes too specific. It can get confused
when it sees new data because it’s only good at recognizing the training
examples.
 W3B: ML Workshop

COMMONLY USED TERMS IN MACHINE LEARNING

Underfitting: Underfitting happens when the computer doesn’t learn
enough from the data. It’s like not studying enough and failing to
understand the problem.
Accuracy: Accuracy is how often the computer gets the right answer.
The higher the accuracy, the better it’s doing!
Confusion Matrix: A confusion matrix is like a table that shows how well
the computer is guessing the labels. It compares the correct answers to
what the model predicted.
Neural Networks: Neural networks are a special type of machine
learning model that is inspired by the human brain. They have layers of
"neurons" that work together to process data.
 W3B: ML Workshop

COMMONLY USED TERMS IN MACHINE LEARNING

Deep Learning: Deep learning is a special type of machine learning that
uses neural networks with many layers. It’s great for really complex
tasks, like understanding pictures or translating languages.
Hyperparameters: Hyperparameters are settings that control how a
machine learning model learns. It’s like choosing the right settings on
your phone or video game to make sure it works just right.
Gradient Descent: Gradient descent is a method used to find the best
solution by making small steps to improve. It’s like climbing a hill and
trying to find the lowest point.
 W3B: ML Workshop

COMMONLY USED TERMS IN MACHINE LEARNING

Clustering: Clustering is when the computer groups things together
based on similarities, without being told what the groups are.
Regression: Regression is a type of machine learning used to predict a
number (like predicting tomorrow’s temperature).
W3B: ML Workshop

TYPES OF MACHINE LEARNING

Supervised Learning:- The model is
trained on labeled data (input-output
pairs).

Unsupervised Learning:- The model

finds patterns in unlabeled data.

Reinforcement Learning:- The model

learns by interacting with the
environment and receiving
rewards/punishments.
W3B: ML Workshop

SUPERVISED VS UNSUPERVISED VS REINFORCEMENT

Feature Supervised Unsupervised Reinforcement

Environment
Data Labelled Data Unlabelled Data
based feedback

Maximise
Find hidden
Goal Predict Outcome Cumulative
pattern
Reward

Market
Example Email Classification Game Playing
Segmentation
 W3B: ML Workshop

SUPERVISED LEARNING AND UNSUPERVISED LEARNING

Supervised Learning:
The model is trained on labeled data (data with input-output pairs).
The goal is to learn a mapping from inputs to outputs, allowing the
model to make predictions on new data.
Supervised Learning Example:
Email Classification: Input: Email text
Output (Label): Spam/Not Spam
Model learns to classify future emails based on
past labeled examples.
 W3B: ML Workshop

SUPERVISED LEARNING AND UNSUPERVISED LEARNING

Unsupervised Learning:
The model is trained on unlabeled data (no explicit output labels).
The goal is to find hidden patterns, structures or groupings within the
data.
Unsupervised Learning Example:
Customer Segmentation: Input: Customer purchase data
No output labels
The model groups similar customers together
(clusters) without predefined categories.
 W3B: ML Workshop
HOW IT WORKS

Supervised Learning Unsupervised Learning

Feed the model input data (features) and Feed the model input data without
correct output labels (targets) labels.

The model adjusts its parameters to

The model looks for patterns, clusters, or
minimize the difference between
anomalies in the data.
predicted and actual outputs.

After training, the model can predict The model learns inherent structures
outputs for unseen data. (e.g., customer groups).
 W3B: ML Workshop
COMMON ALGORITHMS
Supervised Learning Unsupervised Learning

Linear Regression K-Means Clustering

Logistic Regression Principal Component Analysis

Decision Trees Hierarchical Clustering

Support Vector Machines Autoencoders

Neural Networks Gaussian Mixture Models

 W3B: ML Workshop
APPLICATIONS
Supervised Learning Unsupervised Learning

Spam Detection Customer Segmentation

Fraud Detection Market Basket Analysis

Sentiment Analysis Anomaly Detection

Medical Diagnosis Dimensionality Reduction

Credit Scoring Topic Modelling

 W3B: ML Workshop

ADVANTAGES

Supervised Learning:
Highly accurate and interpretable if sufficient labelled data is available.
Provides direct prediction (classification/regression).

Unsupervised Learning:
Handles large amount of unlabeled data which is easier to obtain.
Identifies hidden structures that may not be obvious to humans.
 W3B: ML Workshop

DISADVANTAGES

Supervised Learning:
Labelled data is expensive and time consuming to obtain.
Can overfit if the model is too complex or underfit if it is too simple.

Unsupervised Learning:
Results maybe hard to interpret since there are no explicit labels.
Can produce ambiguous or irrelevant groupings without domain
knowledge.
W3B: ML Workshop

KEY DIFFERENCES

Feature Supervised Unsupervised

Data Type Labelled Data Unlabelled Data

Find hidden patterns or

Goal Predict Outcome
clusters

Clustering/Dimensionality
Task Classification/Regression
Reduction
W3B: ML Workshop

KEY DIFFERENCES

Feature Supervised Unsupervised

Probabilistic (clusters,
Output Definite (direct prediction)
groupings)

Complexity Easier to interpret Harder to interpret

Use Case Known case or labels Discover hidden structures

 W3B: ML Workshop

VISUAL REPRESENTATION
Supervised Learning:
Imagine a scatter plot where points are labelled with colours. The model
learns to seperate those colours.

Unsupervised Learning:
Imagine the same scatter plot without colours. The model identifies
clusters based on the point's proximity to each other.
 W3B: ML Workshop

VISUAL REPRESENTATION
 W3B: ML Workshop

WHEN TO USE
Use Supervised Learning If:
You have labeled data and need to predict an outcome (Example: Fraud
Detection or Price Prediction).

Unsupervised Learning:
You have large amount of data but no labels, and you want to discover
patterns (Example: Customer Segmentation or Market Analysis).
 W3B: ML Workshop

APPLICATIONS OF MACHINE LEARNING

TECH SECTOR
Voice Assistants:- When you say, "Hey Alexa, play my favorite song,"
Alexa listens, understands your voice, and finds the song you like.
Video Recommendations: You know how YouTube suggests videos you
might like? It’s because machine learning studies what you’ve watched
and finds similar ones.
Self-Driving Cars: Cars like Tesla use machine learning to see the road,
notice traffic signs, and even avoid accidents.
 W3B: ML Workshop

APPLICATIONS OF MACHINE LEARNING

HEALTH SECTOR
Detecting Illnesses: Machine learning can look at X-rays or test results
to help doctors find diseases like cancer early.
Fitness Apps: Apps like Fitbit learn about your daily steps and sleep
patterns to give you tips to stay healthy.
Medicine Discovery: It helps scientists find new medicines faster by
studying data about diseases.
 W3B: ML Workshop

APPLICATIONS OF MACHINE LEARNING

MORE AMAZING APPLICATIONS
Education: Apps like Duolingo use machine learning to understand
what you're struggling with and give you personalized lessons.
Games: Machine learning helps make video game characters smarter,
so they feel more real when you play.
Weather Predictions: It studies weather patterns to help predict storms
or sunny days.
 W3B: ML Workshop

THANK YOU!
FOR YOUR ATTENTION