1. Question 1 — What does pooling do?

• ✅ Reduces information while keeping important spatial features

• ❌ Isolates features

• ❌ Combines them

• ❌ Makes them sharper

Explanation:
Pooling downsamples an image but preserves dominant features.

---

2. Question 2 — Max pooling output size

After 2×2 pooling on 26×26 → 13×13.

• ✅ False

• ❌ True

---

3. Question 3 — How do convolutions improve recognition?

• ✅ They capture spatial features

• ❌ Make image clearer

• ❌ Make image smaller

• ❌ Make processing faster

---

4. Question 4 — 28×28 through 3×3 filter

You got this correct.

• ❌ False

• ✅ True

Explanation:
Valid padding 28−3+1 = 26.

---

5. Question 5 — Convolutions’ impact on training

• ❌ Makes it faster

• ❌ No effect

• ✅ Depends on several factors

Explanation:
Training speed depends on hardware, architecture, filter count, image size, etc.

---

🧾 Summary Table

---