IT & Coding Curriculum

Kaushali Club
Block Level Training
 Please
scan and
join the
WhatsAp
p group.
Day 1
Vision of the Kaushali Curriculum

To enable learners to have the skills, knowledge and

mindset to use technology and solve problems that
matter to them, prepare them for the 21st century
and contribute to economic and social progress.
Objectives of the Kaushali Curriculum
Students develop their interest, capability and knowledge
in computing and be able to effectively leverage these
skills to become creators

Students develop computational thinking skills to solve

everyday problems around them

Students develop 21st century skills and mindset to

become responsible digital citizens
Three Days’ Goals
Goals
● Revisit the fundamental concepts from class IX
● Understand the broad areas & themes for class X

● Engage in hands on experience and apply the learnings

● Planning for Further Trainings

A look at the 3 days of training
- Introduction
- Overview of the curriculum of grade IX
- Scratch Basics practice
Day 1
- Problem-solving cycle
- Introduction to grade X curriculum
- Advanced Scratch

- Advanced Scratch (Broadcasting & Sensing)

Day 2 - Fundamentals of AI & Design principles of AI lessons
- Using Generative AI applications
- Introduction to Machine Learning

- Classification and creating a Model (Cassava)

- AI Project Life Cycle
- Training the AI Model - Sameera’s Story
Day 3
- Developing the Model by connecting with Scratch
- Cybersecurity
- Career Options in STEM
 Let’s set the norms for our session
●
together
We pledge to keep our phones on silent mode throughout 4 days of the training
● We pledge to not see the screen of our devices when the session is being delivered.
● We pledge to receive any urgent calls outside of the training hall by silently moving out.
● We pledge to respect each others comment/views/feedback shared during the training.
● We pledge to not participate in any parallel conversation when the presenter is addressing the
entire group.
● We pledge to maintain and abide by the discipline guidelines of the venue.
● We pledge to reach the training hall 5 minutes before the training starts.
● Shuffle the sitting positions daily, sit with people least known to you.
● We pledge for active participation.
● We pledge to ask relevant questions.
● No question is a silly question, feel free to ask any doubt.
● We pledge to keep dining area clean and not leave leftovers On the table
 Now let us introduce ourselves!

ଆପଣଙ୍କ ବ୍ୟାଗ୍ ରେ କ’ଣ ଅଛି (Apanaka bagre kana achhi?)

Remember while introducing yourself you must also share…

…what have you got in your bag for these 4 days?

What we know till
now! (pre-test)
Let’s overview Class IX curriculum
Digital Literacy - Lesson topics and Learning outcomes
Lesson No. Topic Learning Outcomes (students will be able to …)

1 Digital devices ● segregate a group of digital and non digital devices and tell the difference between them

2 Computer networks ● explain internet using the concept of computer networking

3 The World Wide Web ● use search engines to gather useful information from the internet

● write text and adjust attributes like picking a font, font colour, font size, bold, italics,
4 Basics of Word Processing
underline, and alignment

● insert various shapes to a word document like lines, rectangles, basic shapes, callouts,
5 Word processing essentials
etc.

6 Playing around with tables ● organise different type of data in a table (text, numbers, and images)

Project planning, team formation, and

7 ● identify and analyse a local problem using the 5W1H framework for root cause analysis.
ideation

8 Basics of Presentation ● create a new MS Powerpoint presentation, rename it, and open it

9 Infographics and Animation ● use Shapes, Smart Graphics and Tables to demonstrate the obtained information/data

10 Unit Project Presentation ● bring together their learning from the unit in a final project presentation
Data and data analysis - Lesson topics and Learning outcomes

Lesson No. Topic Learning Outcomes (students will be able to …)

1 Data and Excel workbook ● Create and store the data in excel

2 Collecting data and Use of Google Forms ● Employ Google Forms for regular class and school level usage

3 Formatting a spreadsheet ● Develop their understanding of the structure of a spreadsheet

Accessing data resources and defining problem

4 ● Download publicly available data resources for their studies
statement

5 Basic Formulas and Functions in MS-Excel ● Apply the formula for basic data analysis

6 Logical functions ● Apply the formula to create classroom data management

7 Text & other useful functions ● Navigate and manipulate the text data. Also learn commonly used functions

8 Data and Data validation ● Display and pick the specific data for database

9 Presenting data on Chart ● Analyze the data and easily showcase the data

10 Unit Project Presentation ● Use MS-Excel and Data Analysis in meaningful ways related to their daily life
Basic Block based programming - Lesson topics & Learning outcomes

Lesson No. Topic Learning Outcomes (students will learn and explore…)

1 Introduction to Scratch ● get a basic idea of the Scratch tool, Scratch interface and its elements.

2 Basic Concept of Scratch ● sprites, backdrops, and costumes in Scratch.

3 Basic Movements and Animation ● basic movements and animation in Scratch.

4 Events and Sequence ● the concept of sequence and purpose of events blocks in Scratch.

● the concept of loops, control/conditional statements,logical operators and purpose of all

5 Loops, Conditionals and Operators
these blocks in Scratch.

6 Introduction to Problem Solving & Project Ideas ● problem-solving techniques and brainstorm project ideas related to local issues.

7 Planning and Storyboarding ● how to plan and storyboard their Scratch projects.

Project Development: Creating Sprites and ● creating/adding the sprites, backdrops, code and interactions for their group Scratch
8
Backdrops, Coding and Interactions projects.

Project Development: Coding and Interactions,

9 ● creating/adding the coding interactions and test their group Scratch projects.
Finalizing and Testing Projects

10 Unit Project Presentation ● Bring together their learning and present the project developed.
Let’s have some fun with Scratch
Let’s look at a project!

Project link
Let’s understand the project:
1. How many Sprites are there in this project?

2. Which block is used in Scratch to start/run the project?

3. Did you see any Costume blocks used in the project?

4. Were there any Control blocks in the project?

5. What is the backdrop of this project? Can you show me how to change it?
Scratch Interface
Block Based Coding using Scratch

Watch this scratch project and try to experience how CODING

becomes a medium for students to give shape to their
thoughts!

Life Under water

Distance and Displacement
Block Based Coding using Scratch

Time to wear ଶିକ୍ଷାର୍ଥୀର ଟୋପି" (Shiksharthi ra Topi) -

Learner’s Hat

Can you try to create a small project:

- Using at least 4 different kinds of block palette

- Such that you can use it in your subject classroom. E.g create the process of
evaporation, a timeline of major historic events of odisha etc.
Block Based Coding using Scratch

Time for ପ୍ରଦର୍ଶନୀ" (Pradarshani)

- What was the reason behind creating project on this particular topic?

- What challenges you faced while making the project?

- How did you overcome the challenge?

 Fire on the Mountain

Can you please stand up and

move outside to an open space.
Class X curriculum themes
 Informed
Student’s journey in grade 10 career choice
Resilient on STEM
digital citizen
Developing the
AI model

Understanding
AI
Curating fundamentals
project by and AI ethics
Developing using
problem Advance
statement Scratch
A Look at Class 10 Curriculum
Cybersecurity and Machine Learning
Revision of class 9 and creating an ML
its importance in
Scratch model
digital devices

Problem-solving
Social Media: Risks AI and Problem
and Computational
and Reporting Solving and Social
Thinking Artefact
Good

Advanced Scratch: Understanding the Exploring Career

Variables, Operators, Fundamentals of AI Pathways in Coding
Sensing, and and AI
Broadcast
What is your thoughts on this?

Let’s watch the video

Let’s watch a video and identify how they solve a problem!
Share your thoughts…

● What do you think about this video?

● Why do you think they picked this as the problem?

● What steps could you identify for problem solving

process?
What is your thoughts on this?

Problem-
Solving Cycle
What is your thoughts on this?

We understood the problem-solving cycle using

Zipline’s video…
…now let us try to experience this on
our own
Can you relate to it?
Does this picture connect to you
more?
Problem : Heat waves affect students health
Let’s define the problem Statement using RCA
5W1H - a tool for Root Cause Analysis to define the problem statement
Problem statement

Students in classrooms are facing the extreme heat during

class time due to climate change, inadequate ventilation, and
classroom size.
Let’s Think for the Solution
Possible Solutions
Organize health
check up camp
in
the school
Install purified running
water system & giving
ORS to every student.

??

Reminding students to
Sprinkling water on drink water at the
the ground or roof to interval of 45 mins
keep the temperature
low
Let’s find a solution
Possible Solutions
Organize health
check up camp
in
the school
Install purified running
water system & giving
ORS to every student. Shortlisted this
solution as it can be
done quickly, is bold
and is long lasting
Reminding students to
Sprinkling water on drink water at the
the ground or roof to interval of 45 mins
keep the temperature
low
Let’s make a solution
Artifacts : an initial model to test
Anything that communicates what solution you have thought of is an Artifact.
Don’t forget what is our vision!

To enable learners to have the skills, knowledge and

mindset to use technology and solve problems that
matter to them, prepare them for the 21st century and
contribute to economic and social progress.
Let’s Dive Deeper into Scratch
Timer Project:

Project Link
Introduction to Variables :
​ variable is similar to a container that holds information/values that you can
A
access.
Format of Values stored in Variables:
There are three format of values that are usually stored in variables. Mostly used
types of values are​:

● String - contains only alphabets​,alphanumeric

● Integer - contains only number​
● Boolean - either 'True' or 'False'​
Components of Variables:
Declaring a variable in programming has two parts:
1. Variable name: Giving a name to the variable
2. Variable value: Assigning a value to the variable

This variable value can be changed during programming.

Introduction to Variables block in Scratch
Variables palette consists of
the following blocks:
Variable blocks in
● Make a Variable
scratch
● Set
● Change
● Show
● Hide
Creating ‘Timer’ Variable:
Step 1 Step 2 Step 3
Operator Blocks:
Comparison Operators:
● Greater than
● Less than
● Equals to

Math Operators:
● Addition
● Subtraction
● Multiplication
● Division
Hands-on with Scratch
(Timer Project)
Why Variables??
Variables are fundamental components of programming languages and are
used for storing and manipulating data.
● Optimization of code
● Variables enable the reuse of values throughout a program.
● Allows for dynamic behavior
What all we have done today?
Day 2
Ok Google
Use Google Assistant on your phone to -

● Do a google search for the price of a Sony 32 inches TV

● Send a message (SMS or WhatsApp) to a family member or
friend
● Make a call to someone
● Find the directions to Lingaraj Temple
● Find the time it will take to reach Barmunda Bus Stand
● BE CREATIVE!

Say - “Ok Google. Tell me the price of a Sony 32

inches TV.”

Link: assistant.google.com
Lets Recap….
Sensing: Unplugged Activity
Let’s explore type of sense being used by following plants or animals:
Sensing Blocks

Sensing blocks are a type of blocks in Scratch

that allow you to detect and respond to different
factors like touching the sprite/color, draggable
etc.
Usage of Sensing Blocks:
Communication & Mass communication
Communication is the sending and receiving of
information between two persons.

Mass communication refers to a mode of

communication that helps us impart or exchange
information with a large number of people.
Mass Communication Types
Introduction to Broadcast Blocks
● The Broadcast () is an event block.
● The block sends a broadcast messages
throughout the whole scratch project.
● Broadcast in scratch consists of the
following blocks:
Project:

Project link
Scratch Project Presentation by Teachers
Task for School
Make a scratch project using
Variables, Broadcast, Sensing, Text
to speech blocks.
Ice-breaker
Can you give me back the numbers I give you in
reverse order?

Have you tried drawing with Google?

https://quickdraw.withgoogle.com/#details
Training aims
● Develop a knowledge of the AI concepts covered in
the lesson materials
● Become confident that you are equipped to deliver
training sessions and support educators using the
AI lesson resources on an ongoing basis
● To understand the approaches to machine learning
● To increase confidence in using machine learning
for kids to build a machine learning model
 Starter activity

What is intelligence?
Can you describe it?

Commonly, people describe intelligence as the ability to learn and adapt/react to new situations.
Enigma Bombe
Can machines think?
In the 1950s Alan Turing explored the
central question of artificial intelligence
(AI). He thought the original question,
“Can machines think?”

“Can a computer imitate intelligent

behaviour well enough to convince
someone they are talking to a
human?”

Movie recommendation: The

Imitation Game (2014)
Source:
https://en.wikipedia.org/wiki/Alan_Turing#/media/File:Alan_Turing_(191
2-1954)_in_1936_at_Princeton_University.jpg
 CAPTCHA
Commonly Automated Public
Turing test to tell Computers and
Humans Apart
Progress of AI
● General intelligence - competence in a wide
variety of intellectual tasks

AI winter (1970 - 1995) - reduced funding and interest

in AI

● Specialised or narrow intelligence - competence

in a specific intellectual task
Intelligence in machines IS DIFFERENT from intelligence in humans
 Activity 1

Let us try an AI application - ChatGPT

Prompt:

Write a short poem on Odissi

dance.
 Activity 1

Let us try another one.

Prompt:

How many state names in India

start with the letter O?

Chat GPT is an example of Large

language models (LLMs).
 Introduction

Is this piece of paper intelligent?

 Introduction

The ‘intelligent’ piece of paper

My intelligent piece of paper has

never lost a game of tic-tac-toe
(ଛକି - ଶୂନ ଖେଳର).

2 volunteers are needed.

1 will play on behalf of the humans.

1 will follow the instructions given by

the intelligent piece of paper.
 Introduction

What makes the piece of paper “intelligent”?

Move 1:

The piece of paper had instructions

Place an X in a corn
er.
Move 2:
written on it. IF the other player
opposite corner
did not place an O
in the
THEN place an X
in the opposite corn
Does that make the piece of paper 1. er to move
ELSE place an X in
a free corner.
intelligent? Move 3:
IF there are 2 Xs an
d a space in a line
THEN place an X
in the free space on
ELSE IF there are that line.
2 Os and a space in
THEN place an X a line
in that space.
No. ELSE place an X in
a free corner.
Move 4:

The piece of paper is not intelligent.

IF there are 2 Xs an
d a space in a line
THEN place an X
in the free space on
ELSE IF there are that line.
2 Os and a space in
THEN place an X a line
It is a set of IF/THEN rules and
in that space.
ELSE place an X in
a free corner.

cannot adapt itself to new situations. Move 5:

Place an X in the fre
e sp ace.
 Activity 1

Using AI to solve problems

Rule-based Data-driven

Move 1:
Place an X in a corner.

Move 2:
IF the other player did not place an O in the opposite corner
THEN place an X in the opposite corner to move 1.
ELSE place an X in a free corner.

Move 3:
IF there are 2 Xs and a space in a line
THEN place an X in the free space on that line.
ELSE IF there are 2 Os and a space in a line
THEN place an X in that space.
ELSE place an X in a free corner.

Move 4:
IF there are 2 Xs and a space in a line
THEN place an X in the free space on that line.
ELSE IF there are 2 Os and a space in a line
THEN place an X in that space.
ELSE place an X in a free corner.

Move 5:
Place an X in the free space.
 Introduction

Rule based or data driven?

Excel or google sheet that YouTube, Netflix or Jio TV

uses formulas to work out the service that recommends
total cost of a weekly movies based on users’
shopping trip viewing habits

Rule-based Data driven

 Activity 2

What is a model?

A model is a representation of a real-

world context.

For example, as part of your science

projects, you might have made models to
depict processes such as photosynthesis,
the greenhouse effect, and solar-
powered villages.
 Activity 1

What is a data-driven model? Example: Weather forecasting using AI

A data-driven model is used to Training data -> historical weather data

solve a complex problem. (for example, temperature, pressure, wind
speed)
Usually, they rely on a massive
number of examples (lots of data), Predicts -> future weather events
to detect patterns to create the
representation.
 Activity 1

Creating a model
Chess is a complex two-player strategic
board game.

If you wanted to create an AI model of

how to play a game of chess, what would
you need?

Rules

What if we wanted the model to not

just play, but win the game? What
would you need?

Data
 Activity 1

A model of a chess player

What do you want the model to be able to

do?

To play a move! And the next move!

And next.

For each turn, choose which piece to play

and where to move it to have the highest
chance of winning.
 Activity 1

A model of a chess player

What data would you need to create the

model?
● Data of all the chess matches
recorded in history
 Activity 1

Let’s look at this position!

Data needs to be provided to the

machine of all past games
- where this position has been
reached
- which move has been played next,
and
- did that lead to a win, loss, or draw
 Activity 1

AI and chess

1996 was an important year for AI Why is a data-driven approach

development, as an AI system called beneficial in creating an application
Deep Blue famously beat one of the to win a game of chess?
world’s best chess players.
There are more variations of chess games
than there are atoms in the observable
universe.

Picking the best move would be extremely

difficult to achieve with a rule-based
system.
Image source: Encyclopædia Britannica
Anthropomorphisation

To anthropomorphise: “to show or treat an animal,

god, or object as if it is human in appearance,
character, or behaviour”

https://dictionary.cambridge.org/dictionary/english/anthropomorphize

Generated by AI
Anthropomorphisation - we avoid it!

● We avoid using language that: Avoid Instead use

Using phrases such as “AI learns” Use phrases such as “AI

○ Distracts from the understanding or “AI/ML does” applications are designed to” or “AI
developers build applications
that it is people who design and that…”
influence the uses of AI Words that are used to describe the Use system type words (Detect,
behaviour of people (seen, look, input, pattern match, generate,
applications. recognise, create, make) produce)

○ Risks a reduction in the students’

desire to take an active role in
wanting to understand how they
work and be involved in
designing future applications.
AI learning levels framework (SEAME model)
● Activities that relate to the impact of AI on everyday life for society.
Social, Ethical considerations
SE level

Application ● Activities related to developing applications and systems that use AI or

ML models. Example of applications: Chat GPT, Google Gemini
A level

Model ● Concerns the models underlying AI and ML applications.

M level

Engine ● The specific algorithms used to make the models work

E level

(Model adapted from Waite & Curzon, 2018)

Questions?
Using Generative AI Applications
Generative AI applications are built to generate ‘creative’ content
such as:
● Text
● Images
● Sound
● Video
AI applications - generating art
Go to: https://www.craiyon.com/
You will use an application to
generate artwork based on the
criteria you give it.

1. Enter a search term.

2. Select your preferred style
3. Click on Draw.

Generating artwork for ‘sand art’ done on Puri beach

 Activity 2

AI applications generating art task

Use the AI application to generate

Things to think about:
artwork for either:
Who benefits from using AI to
1. A poster on Climate Change,
generate art?
Deforestation or Celebrating
Earth day ● Who owns the art?
● Who made the art? You or the
2. Cover page for a story book or
application?
magazine
What are the problems with using
AI to generate art for you?

99
 Activity 2

Is ChatGPT a Generative AI?

Yes. Chat GPT is an example of LLM or

Large language models (LLMs).

LLMs are a category of models trained on

immense amounts of data making them
capable of understanding and
generating natural language and other
types of content to perform a wide range
of tasks.

10
0
Check point
Questions?
 Activity 3

AI applications — computer vision

Computer vision is a field of AI that

attempts to gain meaningful
information from images.

One way it does this is by identifying

objects in an image.

Why do you think there are

percentages next to each object
that has been identified?

Image source: Shashi Bellamkonda (CC BY 2.0)

10
2
 Activity 3

Computer vision task

Use the website linked on your

Things to think about:
worksheet to choose an image.
Where can you see this technology
It will use AI to predict what is in the
being used and who would benefit
image.
from it and how?
Study the confidence scores.
What might be the problems of
being able to automatically
identify people and objects in
images?

ai-activities.raspberrypi.org/computer-vision-ind/
10
3
 Plenary

AI or not AI?

We will look at the example applications.

For each one, decide whether or not you think it uses AI and justify your answer.

Think about whether it is rule-based or data-driven.

10
4
 Plenary

Application - 1

Excel or google sheet that uses Not AI

formulas to work out the total cost
of a holiday trip to North-East India. The excel or google sheet will work
out the total cost of the holiday
using rule-based formulas.

10
5
 Plenary

Application - 2

YouTube service that recommends AI

videos/songs/films based on users’
viewing habits The service will use the viewing
habits of the user, compare it to
other users, and make a prediction
of what the user would like to watch
next.

There are too many users of an

online streaming service like
YouTube to be able to create rules
for each user.
10
6
 Plenary

Application - 3

Unlocking your smartphones Could be AI

A smartphone uses IF/THEN rules when we
have enabled unlocking method like
specific inputs of PIN number or geometric
patterns to unlock it. This would not use AI.

A face recognition system uses artificial

intelligence to unlock phones. It employs 3D
scanning and machine learning to compare
the user's face with stored information,
ensuring a high level of security.

This is an AI application.

10
7
 Plenary

Summary

AI can be used to solve problems.

AI uses data-driven models to

represent the real world.

Models make predictions.

10
8
Questions and
discussion
Overview of upcoming sessions on AI

● Introduction to machine learning

● Introduction to classification
● Create a machine learning model
● Bias in machine learning models
● Question and answers
 Activity 1

What is machine learning?

Watch the video on YouTube

 Activity 1

How do models ‘learn’?

AI

Machine
The current state learning Not all AI
of AI relies heavily applications use
on the techniques machine learning.
in machine
learning.
 Activity 1

How do models ‘learn’?

Simple problems Complex problems
Use traditional programming, Can’t be solved using traditional
where we can define the exact programming. Require more than a rule-
steps to solve a problem. (rule- based approach.
based approach)

Here, step-by-step instructions We use a lot of data to create machine

are written to solve a problem. learning models. (data-driven approach)

The outputs from a machine learning

model are always predictions. There is
always some uncertainty in the output.
 Activity 1

Would you use machine learning?

Calculating Identifying
Forecasting
the distance animals in a
air pollution
to a star video
 Activity 1

Would you use machine learning?

Calculating Identifying
Forecasting
the distance animals in a
air pollution
to a star video

Probably Could use an Could use an

instructions ML model ML model
 Activity 1

What data would you need?

In order to create a machine learning
model, you need data! Identifying
animals in a
Pick one of the two uses of machine video
learning models. What data would
you need to create a model?

Forecasting
air pollution
 Activity 2

Types of machine learning

Watch the video on YouTube

 Activity 2

Supervised learning
● Needs data that has been
preprocessed by a human Weather
forecasting

● Requires a lot of data to train Computer Chat

Vision GPT
● It’s like a child learning in a
classroom
Generative
AI Google
Quick
Draw
Three types of Machine Learning

Supervised
Learning

Reinforcement Unsupervised
Learning Learning
Questions
What all have we done today?
Day 3
 Plenary

କରିବାର ଅଛି Scratch, ହେଇ ଗଲା କିନ୍ତୁ

Power-cut

12
4
Let’s recap…
Classification
 Activity 3

Classification
Classification is a way of using supervised learning to categorise
data.
A classification model assigns data to a class by applying labels.
 Activity 3

Classification
Classification models must be trained with example data that
already has labels assigned by a human.

What labels would you apply to these images?

 Activity 3

Classification
Classification models must be trained with example data that
already has labels assigned by a human.

What labels would you apply to these images?

 Activity 3

Classification
The more training data you use, the more accurate the model will
be.

Classification
model
 Activity 3

Classification
Once the model is trained, new data can be fed into the model and it
will produce a prediction of which class the data belongs to.
New data

Prediction:
Classification
model Orange
Confidence 92%
 Activity 3

Classification
Class Label

A category that the data Applied to a single piece of data to

can be assigned to indicate which class it belongs to
 Activity 3

Classification — example 1
Computer vision

Using a model to find a range of objects

in a photo or video.

The model is trained with images of

individual objects and of multiple objects,
like this image.
 Activity 3

Classification — example 1
 Activity 3

Classification — example 1
Computer vision

Classes: Shoe, person, dog, bicycle,

wheel, etc.

Benefits: Can identify multiple objects in

an image — useful for technological
advancements such as driverless cars
Classification
checkpoint
Getting hands on
with classification
 Introduction

Machine learning models

We want to create a machine

learning model for a specific
problem.

Recap:

What is a model?

13
8
 Introduction

Using AI to identify disease in crops

A farm wants to modernise its

systems to help produce better
crops. The farm grows cassava and
the owners believe that an AI
application can help them achieve
produce better yields.

They want to use the cameras to

recognise disease in their crops so
they can deal with it before more
damage occurs.

13
9
 Introduction

Cassava

Cassava is also known as Manihot

esculenta, Manioc or Yuca. It is
grown in many countries. It is known
as kanda mula in Odia language.

Cassava root can be boiled and

eaten or used to produce tapioca.

14
0
 Introduction

Vegetable sorting model

A model is a representation of a
real-world context.

The farm has asked you to create a

model that will recognise healthy or
diseased plants, to see if the idea
will work.

What data do you think you will

need?

14
1
 Introduction

Image data

You are about to train your first

machine learning model using
images of healthy and diseased
plants.

You are using training data to train

your model, so which type of
machine learning are you using?
Supervised Unsupervised Reinforcement
learning learning learning

1 2 3
14
2
 Introduction

Image data

You are about to train your first

machine learning model using
images of healthy and diseased
plants.

You are using training data to train

your model, so which type of
machine learning are you using?
Supervised
learning

1
14
3
 Activity 1

Data and machine learning models

Data being used to create an ML model

A set of data is used to create an ML model.

The model is trained with some of the data, Training data

and then tested with the remaining data.

Test data

Once the ML model has been trained and

tested, it is ready to use in the real world.

14
4
 Activity 1

Create your classes

Follow the instructions on the Healthy

worksheet to:

● Create a project that classifies

images
● Create labels for healthy, and
Diseased
diseased plants.
● Add training data
● Train your model

14
5
 Activity 1

Accuracy

The farm has decided that at the

moment, the model is not accurate
enough to consider using.

● Why do you think the farm

thought the model was not
accurate enough?
● What could you do to improve
the accuracy of your model? Recognised as diseased
Think, pair, share with 66% confidence
 Activity 3

Measuring accuracy during testing

There is a simple calculation you can Worked example:

do to measure the accuracy of your
● 13 tests were carried out
model:
● 11 predictions were correct
Accuracy = number of correct predictions
● 11 ÷ 13 = 0.84
. total number
of predictions

This number can be multiplied by ● 0.84 × 100

100 to give the answer as a ● Accuracy = 84%
percentage.
 Activity 3

Accuracy and confidence

ML classification models do not just

return a label, they also return a
confidence score.

A minimum level of confidence

(threshold) should be decided on.

A model might return the correct

label, but we might not be satisfied
with the level of confidence. Recognised as diseased
with 66% confidence
 Activity 3

Accuracy and confidence for our Cassava model

Test 1 Test 2 Test 3 Test 4 Test 5 Test 6 Test 7 Test 8 Test 9 Test 10 Test 11 Test 12 Test 13

Correct predictions Correct Correct Correct Correct Correct Incorrect Correct Incorrect Correct Correct Correct Correct Correct

Confidence score 67% 81% 92% 45% 88% 43% 87% 33% 91% 97% 82% 59% 87%

Confidence Score Outcome of the Model

Case Accuracy
threshold during testing

1 No threshold 11 out of 13 are correct 85%

2 60% threshold 10 out of 13 correct 69%

3 80% threshold 8 out of 13 correct 62%

4 90% threshold 3 out of 13 correct 23%

 Activity 3

Accuracy and confidence for our Cassava model

Test 1 Test 2 Test 3 Test 4 Test 5 Test 6 Test 7 Test 8 Test 9 Test 10 Test 11 Test 12 Test 13

Correct predictions Correct Correct Correct Correct Correct Incorrect Correct Incorrect Correct Correct Correct Correct Correct

Confidence score 67% 81% 92% 45% 88% 43% 87% 33% 91% 97% 82% 61% 87%

Confidence Score Outcome of the Model

Case Accuracy
threshold during testing

1 No threshold 11 out of 13 are correct 85%

2 60% threshold 10 out of 13 correct 77%

3 80% threshold 8 out of 13 correct 62%

4 90% threshold 3 out of 13 correct 23%

 Activity 3

Accuracy and confidence for our Cassava model

Test 1 Test 2 Test 3 Test 4 Test 5 Test 6 Test 7 Test 8 Test 9 Test 10 Test 11 Test 12 Test 13

Correct predictions Correct Correct Correct Correct Correct Incorrect Correct Incorrect Correct Correct Correct Correct Correct

Confidence score 67% 81% 92% 45% 88% 43% 87% 33% 91% 97% 82% 59% 87%

Confidence Score Outcome of the Model

Case Accuracy
threshold during testing

1 No threshold 11 out of 13 are correct 85%

2 60% threshold 10 out of 13 correct 69%

3 80% threshold 8 out of 13 correct 62%

4 90% threshold 3 out of 13 correct 23%

 Activity 3

Accuracy and confidence for our Cassava model

Test 1 Test 2 Test 3 Test 4 Test 5 Test 6 Test 7 Test 8 Test 9 Test 10 Test 11 Test 12 Test 13

Correct predictions Correct Correct Correct Correct Correct Incorrect Correct Incorrect Correct Correct Correct Correct Correct

Confidence score 67% 81% 92% 45% 88% 43% 87% 33% 91% 97% 82% 59% 87%

Confidence Score Outcome of the Model

Case Accuracy
threshold during testing

1 No threshold 11 out of 13 are correct 85%

2 60% threshold 10 out of 13 correct 69%

3 80% threshold 8 out of 13 correct 62%

4 90% threshold 3 out of 13 correct 23%

 Activity 3

Measuring accuracy with a confidence threshold

Accuracy is now the proportion of Worked example:

predictions where:
● 13 tests were carried out
● The label is correct ● 11 predictions were correct, but
the confidence threshold was
and
met for only 8 of these
● The confidence threshold is met
● 8 ÷ 13 = 0.62
● 0.62 × 100
● Accuracy = 62%
 Activity 3

Accuracy and confidence example

What would you set as the

confidence threshold for an ML
model used in an application to help
predict the weather in your area?

● Ideally 100%?
● Would 80% be acceptable?
 Activity 3

Accuracy and confidence example

What would you set as the

confidence threshold for an ML
model used by a driverless car to
identify pedestrians crossing the
road?

● Ideally 100%?
● Would 20% be better?
Creating a Model - Questions
Bias
 Activity 2

Bias

Bias is when the output of a machine learning model favours some

things and deprioritises or excludes others.

Examples of machine learning bias:

Computer programmer jobs only A facial recognition system that is

being shown to users that a social less accurate in recognising people
media platform has identified as with certain skin tones
male
 Activity 2

Data bias

Occurs due to data being used that

does not accurately represent what
is being modelled.

How could data bias appear in your

model for the farmer?

15
9
 Activity 2

Societal bias

This is when data being used to train a What societal bias do you think
model reflects bias that exists in could appear in the data?
society.

Imagine there is an AI application that

predicts the job a person might do.

The model has been trained using

employment data from 1960.
 Activity 2

Societal bias

In the data from 1960, which If gender was in the data used to
gender might have been associated train the model, could this lead to
with the following jobs? the AI application making biased
predictions?
● Cricketer
● Nurse
● Driver
● Teacher
 Activity 2

Data or societal bias?

Computer programmer jobs only Computer Programmers Wanted

being shown to users that a social G Software developers required.

Ideally candidates should be familiar
media platform has identified as DEV with the WebAssembly programming
language, as well as HTML and
male JavaScript.

Societal bias
 Activity 2

Data or societal bias?

A facial recognition system that is

less accurate in recognising people
with certain skin tones

Data bias

Societal bias can lead to data bias

 Plenary

Reducing bias
It is impossible to avoid there being any bias in a data set. However, there are
steps you must take to reduce bias:

● Use large and representative sets of data to train ML models

● Before training the model, separate a proportion of the data that will be
used to create the model for testing
● During the process of designing the ML application, consider people with
different backgrounds, experiences, and opinions
Questions and
discussion
AI Project Lifecycle
 Starter activity

Order the stages of the AI project lifecycle

1 2 3

Evaluating the Explaining Preparing the

model the data
model

4 5 6

Testing the Training the Defining the

model model problem
 Starter activity

Order the stages of the AI project lifecycle

6 3 5 4 1 2

Defining the Preparing the Training the Evaluating the Explaining

Testing the
problem data model model the
model
model
Let’s do an activity!
The Card Exercise

Nobody is allowed to speak during the activity.

Take a minute and form a group.

Breaking Bias & Building
Community
What is Bias? : Bias refers to favours something and
deprioritized or excludes others.

Think on how does the bias makes you feel?

Sameera story (ସମୀରା କାହାଣୀ)
Sameera lives in a small town in India with her
parents and a younger sibling. She is very
curious and ambitious and has so many
dream!

However, at home, her one year younger

Sibling gets most of the attention about his
career and his wishes are prioritized most.

Contd…
Sameera story (ସମୀରା କାହାଣୀ)

But, she has also observed that her parents love her a lot and
care for her too. She tries to communicate this to her parents
but parents generally ignore her.

She talks about it to her friends and notices that they have
similar problems.

Contd…
 Sameera story (ସମୀରା କାହାଣୀ)

She decides to solve the problem and starts on the

problem-solving Journey.

Sameera wants to make her parent understand the biased

statements, words, and references like the picture they use
which are making her feel sad without hurting their feelings.
Can AI Model help Sameera ? (ସମୀରା କିପରି ତା'ର
ପ୍ରୋବ୍ଲେମକୁ ସମାଧାନ କରିବ ? )

● Will Sameera be able to find a solution for

her problem ?
● Can AI make it possible to design a project
to showcase Sameera’s emotion after
listening parents comments?
ଆସନ୍ତୁ ସମୀରାର ପ୍ରୋବ୍ଲେମ ପାଇଁ ଡାଟା ସଂଗ୍ରହ କରିବା
- Let’s collect the data
● What type of words/statements make Sameera happy?
● What type of words/statements feel Sameera sad ?
Example
Happy Sad
Good, Excellent, New ……….. Bad, Poor, Old, Upset, Broken,
Punishment …………
You are so beautiful.
You are a naughty girl.
Congratulation on your achievement.
How dare you to touch my phone!
 Train the AI model
To show Sameera’s Problem, let’s create AI Model using

Machine Learning for Kids Platform

https://drive.google.com/file/d/1sqy-xH_aoTaxKFdG8uSPNRT4JUf6Zo1n/view?usp
=drive_link
Step:1 - Open Machine Learning for Kids(MLK) platform by clicking the given link :
https://machinelearningforkids.co.uk/ and then click on Get started.

Get
Started
Step:2 - Click on Try it now,as we are developing a model without doing any
registration.

If we have already registered

and have a login detail, then we
can click on Log in.

If we want to create a new

log in id , then we can click on
sign up and make a new id.

Try it now
Step:3 - Click on Add a new project to create a new project

Add a new
project
Step:4 - Write Your project name, choose your model type and then select your
language then click on CREATE.

CREATE
Step:5 - Click on your project (Make me Happy) name.
Step:6 - Click on Train option to train your ML model.

Train
Step:7 - Click on Add new label(give name according to your data set in activity
sheet). Here labels means the training data sets for your model. We can add multiple
labels for a single model.

Add new
label
Step:8 - Write the name of your label then click on ADD. To add another label , click
on Add new label.

Add new
label
Step:9 - Click on Add example of two different labels created to train the model.

Add
Add examples
examples
Step:10 - Add more different examples to different labels and then click on Back to
project.

Back to project
Step:11- Click on Learn & Test
Step:12 - Click on Train new machine Learning Model.
Step:13 - After the training is completed then click on Back to project.

Back to project
Step:14 - Here test the accuracy by measuring confidence score .

Test
Confidence Score
Step:15 - Click on Make to make a App in Scratch .
Step:16 - Click on Scratch 3 to redirect to the Scratch interface.
Step:17 - Read the informations given and then click on Open in Scratch 3
Step:18 - Now add the given codes from the respective code block .
Step:19 -Finally,model is ready. Now you can use & test it as per your requirements.
Task for School

https://projects.raspberrypi.org/en/projects/smart-classroom

Explore this project’s pathway and modify with your

idea.
ଚୋରି ପ୍ରତି ସାବଧାନ….!
Let’s understand where we are
● Do you think all the information available
on internet is correct ?

● What all things should be taken care of

while sharing any content online?

● Are all information present on online

resources accurate or legal?

20
0
News on Cyber fraud:
Cybersecurity
● Cyber Security is the protection of computer systems, devices and networks
from any unauthorized access or misuse by others.
● Cybersecurity includes things like secure passwords, internet privacy,captcha etc.
Let’s watch a Video…

Cyber Security Awareness - Think before click - Funny But Serious

Malware and its types
Malware, short for ‘malicious software’,
is software that gets installed on your
laptop, desktop computer or smartphone
and performs a multitude of undesirable
tasks such as stealing passwords, deleting
files or reformatting the hard disk.”
Symptoms of infected systems:
Smartphone Threats
Protection against Smartphone Threats
How much do we know??
● Ever heard of the word “Social Media”, if yes, list some social
media applications that are used on a daily basis?

● Which Social media platforms have you used?

● How many friends have you made on social media?

Social media and its risks
Risk in Social Media

Cyber Grooming

Lets watch a video…..

Risk in Social Media

Cyber Bullying
Reporting Cyber Crime
National Cybercrime Reporting Portal: https://cybercrime.gov.in/ Or Call on : 1930
Resources…..

For further reading, visit https://ciet.ncert.gov.in/ Central Institute of

Educational technology a unit of NCERT.
Impacts of Cyber Crime
If you had not been a teacher…
…then what would you have been?
Let’s play a game!
Implementation at School Level
Consolidation of the Training
Cybersecurity and Machine Learning
Revision of class 9 and creating an ML
its importance in
Scratch model
digital devices

Problem-solving
Social Media: Risks AI and Problem
and Computational
and Reporting Solving and Social
Thinking Artefact
Good

Advanced Scratch: Understanding the Exploring Career

Variables, Operators, Fundamentals of AI Pathways in Coding
Sensing, and and AI
Broadcast
Consolidation: Your feedback
Post training feedback form:
rpf.io/momt