Unit 1.2.1 – Problem Scoping (Revision Time Answers)

1. What are the various stages of AI Project Cycle? Can you explain each with an example?
The AI Project Cycle is a step-by-step process to develop an AI-based solution:

Problem Scoping – Identify a real-world problem and define a clear goal for solving it using AI.
Example: Farmers face pest infestation; goal is to create an AI app to detect pests early.
Data Acquisition – Gather relevant and authentic data related to the problem.
Example: Collect images of pests, farm details, pesticide usage records.
Data Exploration – Visualise and analyse data to identify patterns and trends.
Example: Graph pest counts vs. crop yield to see the correlation.
Modelling – Build or use AI models that can process the data and give predictions.
Example: Train an image recognition model to classify pest types.
Evaluation – Test the AI model using unseen data to measure accuracy and performance.
Example: Compare model’s pest detection results with actual field reports.
Deployment – Implement the solution in the real world.
Example: Deploy the pest detection app on farmers’ smartphones.

2. How is an AI project different from an IT project?

AI Project: Learns from data, improves over time, adapts to changes. Output is based on patterns and predictions (e.g., chatbot that learns from user queries).
IT Project: Works strictly on pre-programmed rules; no learning from new data. Output is fixed unless manually updated (e.g., payroll system).

3. Explain the 4Ws problem canvas in problem scoping.

Who? – Stakeholders directly or indirectly affected by the problem. Example: Farmers facing pest issues.
What? – The nature of the problem and evidence that it exists. Example: Pest infestation in cotton crops confirmed by agricultural reports.
Where? – The location, context, or situation in which the problem occurs. Example: Cotton farms in specific Indian states.
Why? – The benefits of solving the problem and the value for stakeholders. Example: Reduce crop loss, increase farmers’ income.

4. Why is there a need to use a Problem Statement Template during problem scoping?
It summarises the Who, What, Where, and Why in one clear statement. This ensures:

The problem is well-defined and focused.
All stakeholders have a shared understanding.
The project stays aligned to its goal.

5. What is Problem Scoping? What are the steps of Problem Scoping?
Problem Scoping is the process of identifying a problem worth solving with AI, understanding its context, and setting a clear goal.
Steps:

Select a theme (e.g., Health, Agriculture, Education).
Choose a topic under that theme (e.g., Pest control in cotton farming).
List specific problems in that topic.
Select one problem to solve.
Apply the 4Ws canvas.
Write the problem statement template.

6. Who are the stakeholders in the problem scoping stage?
Stakeholders are individuals, groups, or organisations affected by the problem or who benefit from the solution. Example: Farmers, agricultural officers, pesticide suppliers.

Unit 1.2.2 – Data Acquisition (Revision Time Answers)

1. How will you differentiate between Training Data and Testing Data? Elaborate with examples.

Training Data: Used to teach the AI model patterns and relationships.
Example: Previous pest images and details to train a pest detection model.
Testing Data: Unseen data used to check the model’s performance.
Example: New pest images not used during training, to check if the model predicts correctly.

2. Name various methods for collecting data. For each method, can you name at least one project in which you may use that method of data collection?

Surveys – Collecting opinions/feedback; Example: Customer satisfaction survey for a retail AI chatbot.
Sensors – Collecting real-time environmental data; Example: Temperature sensors for predicting crop diseases.
Web Scraping – Collecting data from websites; Example: Scraping product prices for an AI price-comparison tool.
Observation/Field Data Collection – On-site manual data gathering; Example: Photographing pests in cotton fields.
Government Open-Data Portals – Accessing public datasets; Example: Using data.gov.in for traffic and weather data.

3. What must you keep in mind while collecting data so it is useful?

Data must be relevant to the problem.
Data must be authentic and from reliable sources.
Data should be open-source and legally usable.
The dataset should be complete, with minimal missing or duplicate entries.

5. Name a few government websites from where you can get open-source data.

data.gov.in – Government of India’s open data portal.
india.gov.in – National portal for official datasets and information.

REVISION TIME:1.2.3

Q1. What is the significance of Data Exploration after you have acquired the data for the problem scoped? Explain with examples.

Answer:
Data Exploration is significant because it helps in understanding the quality and nature of the data collected. It helps find trends and patterns out of the data which would help you develop a strategy for your project. It detects missing values, outliers, inconsistencies, and hidden patterns that can influence the accuracy of a model.
Example: In a student marks dataset, exploring the data might reveal missing marks for some students or unusually high/low scores. Identifying these issues ensures that the model built later is reliable.

Q2. What do you think is the relevance of Data Visualization in AI?

Answer:
Data Visualization makes complex data easy to understand by presenting it in the form of charts and graphs. In AI, it helps in:

Spotting patterns and relationships between variables,
Detecting anomalies and trends,
Communicating results clearly to stakeholders.
It bridges the gap between raw data and decision-making.

Q3. List any five graphs used for data visualization.

Answer:

Bar Graph
Histogram
Line Chart
Scatter Plot
Pie Chart

Q4. How is Data Exploration different from Data Acquisition?

Answer:

Data Acquisition is the process of collecting relevant data from sources (e.g., surveys).
Data Exploration is the process of analysing and visualising the collected data to identify errors, patterns, and insights.

In short: Acquisition = Collecting data, Exploration = Understanding data.

Q5. Use an example to explain at least one Data Visualization technique.

Answer:
Example — Histogram:
A histogram is used to show the frequency distribution of numerical data.
If we plot a histogram of students’ marks in Mathematics, the x-axis shows mark ranges (0–10, 11–20, …, 91–100) and the y-axis shows the number of students in each range. This quickly shows how marks are distributed — whether most students scored average, low, or high.

REVISION TIME:1.2.4

Q1. What are the various stages of the AI Project Cycle? Explain each with examples.

Answer:
The AI Project Cycle has six stages:

Problem Scoping – Define the problem clearly using 4Ws (Who, What, Where, Why).
Example: Farmers face crop pest attacks — problem scoped as “detect pests early.”
Data Acquisition – Collect relevant data from sources like surveys, sensors, open datasets.
Example: Gather images of infected and healthy crops.
Data Exploration – Analyse and visualise the collected data for patterns, missing values, or biases.
Example: Use scatter plots to see how pest images differ.
Modelling – Train an AI model using the prepared data.
Example: Build an image classifier to identify pests.
Evaluation – Test the model on unseen data and measure metrics (accuracy, precision, recall, F1-score).
Example: Check if the pest classifier correctly predicts infected crops.
Deployment – Integrate the model into real-world usage.
Example: Release a mobile app for farmers to click crop photos and get pest detection results.

Q2. What is Artificial Intelligence? Give an example where AI is used in day-to-day life.

Answer:
Artificial Intelligence (AI) is the ability of a machine to perform tasks that normally require human intelligence, such as learning, reasoning, and decision-making.

Examples in daily life:

Google Maps suggesting fastest routes,
Virtual assistants like Alexa or Siri,
Spam filtering in emails.

Q3. How is Machine Learning related to Artificial Intelligence?

Answer:
Machine Learning (ML) is a subset of Artificial Intelligence.

AI is the broad field of creating intelligent machines.
ML gives machines the ability to learn patterns from data without explicit programming.

Example: AI = self-driving car; ML = algorithm inside it that learns from past driving data to detect pedestrians.

Q4. Compare and contrast Rule-based and Learning-based approach in AI modelling indicating clearly when each of these may be used.

Answer:

Rule-based approach:
- Works on fixed “if–then” rules defined by humans.
- Suitable when rules are simple and well-defined.
- Example: If temperature > 37°C then “Fever”.
Learning-based approach:
- Learns automatically from large datasets using algorithms.
- Suitable for complex tasks where rules are not easily defined.
- Example: Recognising human faces in photos.

Comparison:

Rule-based is explicit and simple but fails with complexity.
Learning-based adapts to new data and complex patterns but requires training data.

REVISION TIME:1.2.5

Q1. What is Evaluation?

Answer:
Evaluation is the stage of the AI Project Cycle where the performance of the trained model is tested on unseen data. It checks how accurately the model makes predictions using metrics such as accuracy, precision, recall, and F1-score.

Q2. What are various Model evaluation techniques?

Answer:
Common model evaluation techniques include:

Confusion Matrix – compares actual and predicted values.
Accuracy – ratio of correct predictions to total predictions.
Precision – ratio of true positives to all predicted positives.
Recall – ratio of true positives to all actual positives.
F1-score – harmonic mean of precision and recall.

Q3. Why is model evaluation important in AI projects?

Answer:
Model evaluation is important because:

It ensures the model is reliable and accurate.
It helps identify errors such as false positives and false negatives.
It checks whether the model generalises well to real-world data.
It prevents deploying a model that may be biased or ineffective.

Q4. What do you understand by the terms True Positive and False Positive?

Answer:

True Positive (TP): When the model correctly predicts a positive case.
Example: Predicting a patient has a disease when they actually have it.
False Positive (FP): When the model incorrectly predicts positive for a case that is actually negative.
Example: Predicting a healthy patient has a disease.

REVISION TIME:1.2.6

Q1. Rearrange the steps of AI Project Cycle in correct order:

a. Data Acquisition
b. Problem Scoping
c. Modelling
d. Data Exploration
e. Deployment
f. Evaluation

Answer (Correct Order):
b → a → d → c → f → e
(Problem Scoping → Data Acquisition → Data Exploration → Modelling → Evaluation → Deployment)

Q2. The process of breaking down the big problem into a series of simple steps is known as:

a. Efficiency
b. Modularity
c. Both a) and b)
d. None of the above

Answer: b. Modularity

Q3. The primary purpose of data exploration in AI project cycle is:

a. To make data more complicated
b. To simplify complex data
c. To discover patterns and insights in data
d. To visualize data

Answer: c. To discover patterns and insights in data

Q4. Deployment is the final stage in the AI project cycle where the AI model or solution is implemented in a real-world scenario. (True/False)

Answer: True

Q5. Identify A, B and C in the following diagram (Hint: How AI, ML & DL are related):

(The handbook shows three concentric circles — AI → ML → DL)

Answer:

A = Artificial Intelligence (AI)
B = Machine Learning (ML)
C = Deep Learning (DL)

AI Ethics – Section 1.3

1. The guiding principles to decide what is good or bad is known as

Answer: Ethics

2. When building AI solutions, we need to ensure that they follow

Answer: AI Principles

3. Praneet has taken extra packets of mouth freshener after dinner from a restaurant. Is it considered as theft? Is it Moral or Ethical concern?

Answer: No, it is not considered theft. It is an Ethical concern.

4. Rakshit and Aman are talking about purchasing a new mobile. Aman starts getting notifications of various models of mobiles that meet his requirement. Which ethical concern does this depict?

Answer: Data Privacy

5. “Preference for one over the other” is known as

Answer: Bias

6. Artificial Intelligence and machine learning systems can display unfair behaviour if not trained properly.

Answer: True

7. Search for images of personal secretary on Google, displaying predominantly the images of women is an example of

Answer: Bias

8. An Ethical AI framework makes sure that transparency, fairness and accountability is develop into the systems to provide unbiased results.

Answer: True

ANSWER THE FOLLOWING:

Q1. Differentiate between Ethics and Morals with suitable examples

Answer: Morals: The beliefs dictated by our society. They are not fixed and can be different for different societies. Examples: Always speak the truth, Always be loyal, Always be generous.

Ethics: The guiding principles to decide what is good or bad. These are values that a person themselves chooses for their life. Examples: Is it good to speak the truth in all situations? Is it good to be loyal under all circumstances?

Q2. Define Principles of AI

The following principles in AI Ethics affect the quality of AI solutions:
1. Human Rights – AI should be designed to respect fundamental human rights, ensuring dignity, freedom, and equality for all individuals.
2. Bias – Bias in AI refers to preference for one over the other which can lead to unfair treatment of individuals or groups when systems are not trained properly.
3. Privacy – Data privacy, referred to as information privacy, is concerned with the proper handling of sensitive data including personal and confidential data, protecting its confidentiality and immutability.
4. Inclusion – AI must not discriminate against a particular group of population causing them any kind of disadvantage. It should include and benefit all sections of society equally.

Q3. Explain Data Privacy

Answer: Data privacy, referred to as information privacy, is concerned with the proper handling of sensitive data including personal data and other confidential data, to meet regulatory requirements as well as protecting the confidentiality and immutability of the data.

Q4. Craft a description of how considerations for inclusivity are addressed during the development of AI models

Answer: AI must not discriminate against a particular group of population, causing them any kind of disadvantage. We must check: Does your AI leave out any person or a group? Is a rich person and a poor person benefitted equally from your AI? How easy is it to use your AI? Who does your AI help?
Q5. Write Major Issues around AI Ethics

Answer:
• Collection and misuse of personal data.
• Violation of data privacy.
• AI influencing human decision-making in unfair ways.
• Bias in AI leading to unfair treatment of certain groups.
• AI depriving people of opportunities.
• Issues of transparency, fairness, and accountability in AI solutions.

Q6. A company had been working on a secret AI recruiting tool which penalised resumes of women chefs.
a. What aspect of AI ethics is illustrated?
b. What could be the possible reasons?

Answer:

a. Aspect illustrated: Bias.
b. Possible reasons: The bias in training data taught the system that male candidates are preferable, leading to unfair discrimination against resumes with ‘women chef’.

Q7. AI replacing humans in factories – Ethical Concern. What according to you is a better approach? Justify.

Answer: A better approach is to see AI as meant to ease our lives. If machines overtake monotonous and tedious tasks, humans should upgrade their skills to remain their masters always.