How training data shapes AI behavior in AI for Everyone - Practice Exercises

Challenge - 5 Problems

🧠 Conceptual

intermediate

How does biased training data affect AI predictions?

Imagine an AI model trained mostly on images of cats with white fur. What is the most likely effect on the AI's ability to recognize cats with black fur?

A The AI will have difficulty recognizing black cats because it saw mostly white cats during training.

B The AI will ignore fur color and only focus on cat shapes, so no effect.

C The AI will recognize black cats just as well as white cats.

D The AI will refuse to classify any cat that is not white.

💡 Hint

Think about what the AI learned from the examples it saw most often.

If the training data mostly shows white cats, the AI learns features common to white cats. It may not learn features of black cats well, so it struggles to recognize them.

🔍 Analysis

intermediate

Output of model accuracy with imbalanced training data

Consider a classification model trained on 95% class A and 5% class B data. After training, the model predicts all inputs as class A. What is the approximate accuracy on the training set?

AI for Everyone

total_samples = 1000
class_A_samples = 950
class_B_samples = 50
correct_predictions = 950  # model predicts all as class A
accuracy = correct_predictions / total_samples
print(f"Accuracy: {accuracy:.2f}")

A Accuracy: 0.05

B Accuracy: 0.95

C Accuracy: 1.00

D Accuracy: 0.50

💡 Hint

How many samples are correctly predicted if the model always predicts class A?

The model correctly predicts all 950 class A samples but misses all 50 class B samples. Accuracy = 950/1000 = 0.95.

❓ Hyperparameter

advanced

Choosing training data size for better AI behavior

You want to improve an AI model's ability to generalize to new data. Which training data strategy is best?

A Use a large, diverse dataset covering many scenarios.

B Use a very small, clean dataset to avoid noise.

C Use only data from one specific case to specialize the model.

D Use random data unrelated to the task to increase size.

💡 Hint

Think about what helps the model learn to handle different situations.

A large and diverse dataset helps the model learn many patterns and generalize well to new data. Small or narrow data limits learning, and unrelated data confuses the model.

❓ Metrics

advanced

Interpreting precision and recall with skewed training data

An AI model trained on data with 90% negative and 10% positive cases achieves 90% accuracy. However, its precision is 0.5 and recall is 0.1 for the positive class. What does this tell you about the model's behavior?

A The model is good at finding positive cases but often makes false alarms.

B The model ignores the positive class and only predicts negatives.

C The model perfectly identifies positive cases but fails on negatives.

D The model misses many positive cases and is not reliable at identifying them.

💡 Hint

Recall measures how many actual positives are found. Precision measures how many predicted positives are correct.

Low recall (0.1) means the model finds only 10% of actual positives. Precision 0.5 means half of predicted positives are correct. So the model misses many positives and is unreliable for that class.

🔍 Analysis

expert

Why does this AI model fail to learn from training data?

Given this training loop snippet, why does the model's loss not decrease?

for epoch in range(5):
    for x, y in train_loader:
        optimizer.zero_grad()
        output = model(x)
        loss = loss_fn(output, y)
        loss.backward()
        optimizer.step()

A The model(x) call is missing input data, so output is empty.

B The optimizer.step() is called before loss.backward(), so gradients are not computed.

C The loss.backward function is not called properly because parentheses are missing.

D The optimizer.zero_grad() is called after loss.backward(), so gradients accumulate.

💡 Hint

Check how functions are called in Python.

loss.backward is a function and must be called with parentheses: loss.backward(). Without parentheses, gradients are not computed, so optimizer.step() does nothing.

Practice

1. Why is training data important for AI behavior?

easy

A. Because training data controls the hardware AI runs on

B. Because training data is used to write AI's code manually

C. Because training data is only used after AI is fully built

D. Because AI learns patterns from training data to make decisions

Solution

1. Step 1: Understand AI learning process

   AI learns by finding patterns in examples called training data.

2. Step 2: Connect training data to AI decisions

   Good training data helps AI make better decisions by showing it what to expect.

3. Final Answer:

   Because AI learns patterns from training data to make decisions -> Option D

4. Quick Check:

   Training data shapes AI behavior = C [OK]

Hint: AI learns from examples, not from code or hardware [OK]

Common Mistakes:

• Thinking AI code is written from training data
• Confusing training data with hardware
• Believing training data is used after AI is built

2. Which of these is the correct way to describe training data's role?

easy

A. Training data is the set of examples AI uses to learn

B. Training data is the AI's final output after learning

C. Training data is the software that runs AI models

D. Training data is the hardware AI uses to process information

Solution

1. Step 1: Define training data

   Training data is the collection of examples given to AI to learn from.

2. Step 2: Eliminate incorrect options

   Training data is not output, software, or hardware but input examples.

3. Final Answer:

   Training data is the set of examples AI uses to learn -> Option A

4. Quick Check:

   Training data = learning examples [OK]

Hint: Training data = examples AI learns from, not output or hardware [OK]

Common Mistakes:

• Confusing training data with AI output
• Mixing training data with software or hardware
• Thinking training data runs the AI

3. Consider this code snippet simulating AI training data influence:

training_data = ['cat', 'dog', 'cat', 'bird']
model_behavior = {animal: training_data.count(animal) for animal in set(training_data)}
print(model_behavior)

What is the output?

medium

A. {'cat': 1, 'dog': 1, 'bird': 1}

B. {'cat': 2, 'dog': 1, 'bird': 1}

C. {'cat': 3, 'dog': 1, 'bird': 1}

D. Error: count method not found

Solution

1. Step 1: Understand the code logic

   The code counts how many times each unique animal appears in training_data.

2. Step 2: Count occurrences

   'cat' appears twice, 'dog' once, 'bird' once, so the dictionary reflects these counts.

3. Final Answer:

   {'cat': 2, 'dog': 1, 'bird': 1} -> Option B

4. Quick Check:

   Count of each animal in list = {'cat': 2, 'dog': 1, 'bird': 1} [OK]

Hint: Count items in list using count() for each unique element [OK]

Common Mistakes:

• Assuming count counts total list length
• Confusing set with list order
• Thinking count method causes error

4. This code tries to count training data items but has a bug:

training_data = ['apple', 'banana', 'apple']
model_behavior = {animal: training_data.count for animal in set(training_data)}
print(model_behavior)

What is the error and how to fix it?

medium

A. Error: set() cannot be used on lists

B. Error: print() cannot display dictionaries

C. Error: count is a method, should be called with parentheses like count()

D. No error, code runs fine

Solution

1. Step 1: Identify the bug in code

   The code uses training_data.count without parentheses, so it refers to the method itself, not calling it.

2. Step 2: Fix the method call

   Adding parentheses and argument fixes it: training_data.count(animal)

3. Final Answer:

   Error: count is a method, should be called with parentheses like count() -> Option C

4. Quick Check:

   Method call missing parentheses = B [OK]

Hint: Remember to call methods with () to execute them [OK]

Common Mistakes:

• Thinking set() can't be used on lists
• Believing print can't show dictionaries
• Ignoring missing parentheses on method calls

5. An AI model trained only on photos of cats and dogs struggles to recognize birds. What is the best way to improve its behavior?

hard

A. Add more bird photos to the training data

B. Remove all cat and dog photos from training data

C. Change the AI model architecture without changing data

D. Train the AI only on bird photos

Solution

1. Step 1: Identify the problem with training data

   The model lacks examples of birds, so it cannot learn to recognize them well.

2. Step 2: Improve training data quality

   Adding bird photos gives the AI examples to learn bird features, improving recognition.

3. Final Answer:

   Add more bird photos to the training data -> Option A

4. Quick Check:

   More diverse training data = better AI behavior [OK]

Hint: Add missing examples to training data to improve AI [OK]

Common Mistakes:

• Removing existing data reduces AI knowledge
• Changing model without data won't fix missing knowledge
• Training only on birds loses knowledge of cats and dogs