1. Question 1

• ❌ Multiply weights

• ❌ Linearly separate data

• ✅ Introduce non-linearity by mapping inputs to [0, 1]

• ❌ Map to [-1, 1]

---

2. Question 2

• ❌ One hidden + output only

• ❌ Two hidden layers

• ❌ Two input layers

• ✅ One input layer and one output layer

---

3. Question 3

• ❌ Reduces parameters

• ❌ Causes underfitting

• ❌ Decreases flexibility

• ✅ Increases model flexibility

---

4. Question 4

• ❌ Add more layers

• ❌ Adjust weights

• ❌ Shift decision boundary

• ✅ Apply a different activation function

---

5. Question 5

• ❌ Too many neurons

• ❌ Too few neurons

• ❌ High learning rate

• ✅ Insufficient training data

---

6. Question 6

• ❌ Captured all patterns

• ✅ Model cannot capture data complexity

• ❌ Too many layers

• ❌ Too complex

---

7. Question 7

• ❌ Add hidden layers

• ✅ Increase number of output neurons = number of classes

• ❌ Use single neuron

• ❌ Use sigmoid

---

8. Question 8

• ❌ Mean Squared Error

• ❌ Hinge Loss

• ❌ Binary Cross Entropy

• ✅ Cross Entropy

---

9. Question 9

• ✅ Compute gradient for updating weights

• ❌ Forward propagation

• ❌ Reduce layers

• ❌ Apply activation

---

10. Question 10

• ❌ Softmax

• ❌ Sigmoid

• ❌ Tanh

• ✅ ReLU

---

🧾 Summary Table

---