Data Science and AI Familiarisation Activity
Question 1 — Applications of AI/ML in My Area of Specialisation
In my field of computer science and data science, AI and ML are already transforming the
way we work. Some examples include: - Chatbots and language translators that
understand and respond in natural language. - Image recognition systems that can detect
diseases in X-rays or recognise faces in photos. - Predictive tools that forecast the
weather, stock prices, or even product demand. - Recommendation engines that suggest
movies, courses, or products based on our past choices. - Autonomous systems like
self-driving cars and drones that make decisions on their own.
Question 2 — How Children Learn to Differentiate Between Dogs
and Cats
Babies can start telling the difference between dogs and cats as early as 3–4 months old
just by looking at shapes and patterns. By the time they are 2–3 years old, they can do this
confidently. They learn through: - Seeing the animals often — in real life, in books, or on
screens. - Listening — a bark means “dog,” a meow means “cat.” - Being corrected gently
by parents or others when they make a mistake. - Recognising key features like ear
shape, body size, and tail type.
Question 3 — Spam Messages and Their Patterns
Examples of spam messages: 1. “Congratulations! You’ve won ■10,00,000. Click here to
claim your prize.” 2. “Your bank account will be blocked today! Verify your details now.” 3.
“Earn ■50,000 per week from home — No experience required.” 4. “You have an
unclaimed lottery reward. Confirm your identity here.” 5. “Flash sale! 90% off branded
watches — Today only!” Common patterns: - They create urgency or a sense of fear. -
They promise unrealistic rewards. - They ask for personal or bank details. - They often
include suspicious or shortened links. - The grammar and punctuation are often strange.
Question 4 — Features for Loan Sanction and AI Automation
Banks usually check these things before approving a loan: - Credit score and repayment
history. - Monthly income and how stable your job is. - How much debt you already have
compared to your income. - Your age, education, and how long you’ve lived at your
address. - The value of any collateral you can provide. - Patterns in your bank
transactions. How AI can help: A machine learning model can take all this information and
give a “risk score” for each applicant. This can help bank staff make faster and fairer
decisions. The system can also explain why a loan was approved or rejected, so it stays
transparent.
Question 5 — MNIST Dataset Exploration
1. Total images: 70,000 in total — 60,000 for training and 10,000 for testing. 2.
Observations for digit “0”: Most are circular or oval, but there are differences in thickness,
size, and neatness. Some are slightly tilted or open. 3. Digit 5 at sample index 2258: Looks
more like a “6” because the lower curve is closed. 4. Similarities and differences in strokes:
- 0 and 8 both have loops, but 8 has two stacked loops. - 1 and 7 are both
straight-line-based, but 7 has a top bar. - 6 and 9 have similar loops but in opposite
directions. 5. Challenges for a machine: - Everyone’s handwriting is different. - Some digits
are messy or unclear. - Certain numbers look very similar to others. - Images can have
noise or imperfections. 6. Where it’s useful: - Reading postal codes. - Processing bank
cheques. - Digitising forms. - Recognising vehicle number plates.
Question 6 — Height & Weight Data
Height (cm) Weight (kg)
150 45
155 50
160 54
165 59
170 63
175 70
180 75
Observation: The scatter plot shows a positive correlation — taller students usually weigh
more, though there are some exceptions.