AI Neural Networks Basics MCQ Questions with Answers (Latest 2026)

Q1. Which option best describes a perceptron?

A. A clustering algorithm. B. Linear classifier with a step activation. C. A pure regression model. D. A decision tree.

Select an answer to check.

Answer: Linear classifier with a step activation.

Here, Linear classifier with a step activation. is the right choice. Historic building block. It aligns directly with what the question asks about which option best describes a perceptron. A quick elimination of partially true options helps confirm it.

Q2. What is the primary purpose of a perceptron?

A. A clustering algorithm. B. A decision tree. C. A pure regression model. D. Linear classifier with a step activation.

Select an answer to check.

Answer: Linear classifier with a step activation.

In this case, Linear classifier with a step activation. is correct. Historic building block. It aligns directly with what the question asks about what is the primary purpose of a perceptron. A quick elimination of partially true options helps confirm it.

Q3. Which statement about a perceptron is most accurate?

A. A decision tree. B. A clustering algorithm. C. Linear classifier with a step activation. D. A pure regression model.

Select an answer to check.

Answer: Linear classifier with a step activation.

The best option here is Linear classifier with a step activation.. Historic building block. It aligns directly with what the question asks about which statement about a perceptron is most accurate. A quick elimination of partially true options helps confirm it.

Q4. How is a perceptron best characterized?

A. Linear classifier with a step activation. B. A pure regression model. C. A decision tree. D. A clustering algorithm.

Select an answer to check.

Answer: Linear classifier with a step activation.

For this question, Linear classifier with a step activation. is correct. Historic building block. It aligns directly with what the question asks about how is a perceptron best characterized. A quick elimination of partially true options helps confirm it.

Q5. Which option best describes an MLP?

A. A pure CNN. B. A single linear layer only. C. Multi-layer perceptron; stack of dense layers with nonlinearity. D. A pure transformer.

Select an answer to check.

Answer: Multi-layer perceptron; stack of dense layers with nonlinearity.

Multi-layer perceptron; stack of dense layers with nonlinearity. is the correct answer here. Universal approximator. It aligns directly with what the question asks about which option best describes an mlp. A quick elimination of partially true options helps confirm it.

Q6. What is the primary purpose of an MLP?

A. A single linear layer only. B. A pure CNN. C. A pure transformer. D. Multi-layer perceptron; stack of dense layers with nonlinearity.

Select an answer to check.

Answer: Multi-layer perceptron; stack of dense layers with nonlinearity.

Here, Multi-layer perceptron; stack of dense layers with nonlinearity. is the right choice. Universal approximator. This matches the core idea being tested around what is the primary purpose of an mlp. A quick elimination of partially true options helps confirm it.

Q7. Which statement about an MLP is most accurate?

A. A pure CNN. B. A pure transformer. C. A single linear layer only. D. Multi-layer perceptron; stack of dense layers with nonlinearity.

Select an answer to check.

Answer: Multi-layer perceptron; stack of dense layers with nonlinearity.

In this case, Multi-layer perceptron; stack of dense layers with nonlinearity. is correct. Universal approximator. This matches the core idea being tested around which statement about an mlp is most accurate. A quick elimination of partially true options helps confirm it.

Q8. How is an MLP best characterized?

A. A pure transformer. B. A single linear layer only. C. A pure CNN. D. Multi-layer perceptron; stack of dense layers with nonlinearity.

Select an answer to check.

Answer: Multi-layer perceptron; stack of dense layers with nonlinearity.

The best option here is Multi-layer perceptron; stack of dense layers with nonlinearity.. Universal approximator. This matches the core idea being tested around how is an mlp best characterized. A quick elimination of partially true options helps confirm it.

Q9. Which option best describes a hidden layer?

A. A non-input/non-output layer in a neural net. B. The loss function. C. The output layer only. D. The optimizer.

Select an answer to check.

Answer: A non-input/non-output layer in a neural net.

For this question, A non-input/non-output layer in a neural net. is correct. Multiple hidden layers = deep. This matches the core idea being tested around which option best describes a hidden layer. A quick elimination of partially true options helps confirm it.

Q10. What is the primary purpose of a hidden layer?

A. The optimizer. B. The output layer only. C. A non-input/non-output layer in a neural net. D. The loss function.

Select an answer to check.

Answer: A non-input/non-output layer in a neural net.

A non-input/non-output layer in a neural net. is the correct answer here. Multiple hidden layers = deep. This matches the core idea being tested around what is the primary purpose of a hidden. A quick elimination of partially true options helps confirm it.

Q11. Which statement about a hidden layer is most accurate?

A. The loss function. B. The output layer only. C. The optimizer. D. A non-input/non-output layer in a neural net.

Select an answer to check.

Answer: A non-input/non-output layer in a neural net.

Here, A non-input/non-output layer in a neural net. is the right choice. Multiple hidden layers = deep. That is exactly the concept behind which statement about a hidden layer is most in this context. A quick elimination of partially true options helps confirm it.

Q12. How is a hidden layer best characterized?

A. The loss function. B. The optimizer. C. The output layer only. D. A non-input/non-output layer in a neural net.

Select an answer to check.

Answer: A non-input/non-output layer in a neural net.

In this case, A non-input/non-output layer in a neural net. is correct. Multiple hidden layers = deep. That is exactly the concept behind how is a hidden layer best characterized in this context. A quick elimination of partially true options helps confirm it.

Q13. Which option best describes a weight?

A. A hyperparameter. B. A label. C. An activation value. D. A learned coefficient on an input/connection.

Select an answer to check.

Answer: A learned coefficient on an input/connection.

The best option here is A learned coefficient on an input/connection.. Trained via gradient descent. That is exactly the concept behind which option best describes a weight in this context. A quick elimination of partially true options helps confirm it.

Q14. What is the primary purpose of a weight?

A. An activation value. B. A label. C. A hyperparameter. D. A learned coefficient on an input/connection.

Select an answer to check.

Answer: A learned coefficient on an input/connection.

For this question, A learned coefficient on an input/connection. is correct. Trained via gradient descent. That is exactly the concept behind what is the primary purpose of a weight in this context. A quick elimination of partially true options helps confirm it.

Q15. Which statement about a weight is most accurate?

A. An activation value. B. A hyperparameter. C. A learned coefficient on an input/connection. D. A label.

Select an answer to check.

Answer: A learned coefficient on an input/connection.

A learned coefficient on an input/connection. is the correct answer here. Trained via gradient descent. That is exactly the concept behind which statement about a weight is most accurate in this context. A quick elimination of partially true options helps confirm it.

Q16. How is a weight best characterized?

A. A hyperparameter. B. An activation value. C. A label. D. A learned coefficient on an input/connection.

Select an answer to check.

Answer: A learned coefficient on an input/connection.

Here, A learned coefficient on an input/connection. is the right choice. Trained via gradient descent. It fits the requirement in the prompt about how is a weight best characterized. A quick elimination of partially true options helps confirm it.

Q17. Which option best describes a bias term?

A. A weight on an input. B. A learning rate. C. Per-neuron additive shift. D. A regularization penalty.

Select an answer to check.

Answer: Per-neuron additive shift.

In this case, Per-neuron additive shift. is correct. Adds expressiveness. It fits the requirement in the prompt about which option best describes a bias term. A quick elimination of partially true options helps confirm it.

Q18. What is the primary purpose of a bias term?

A. Per-neuron additive shift. B. A learning rate. C. A regularization penalty. D. A weight on an input.

Select an answer to check.

Answer: Per-neuron additive shift.

The best option here is Per-neuron additive shift.. Adds expressiveness. It fits the requirement in the prompt about what is the primary purpose of a bias. A quick elimination of partially true options helps confirm it.

Q19. Which statement about a bias term is most accurate?

A. A weight on an input. B. A learning rate. C. A regularization penalty. D. Per-neuron additive shift.

Select an answer to check.

Answer: Per-neuron additive shift.

For this question, Per-neuron additive shift. is correct. Adds expressiveness. It fits the requirement in the prompt about which statement about a bias term is most. A quick elimination of partially true options helps confirm it.

Q20. How is a bias term best characterized?

A. A weight on an input. B. Per-neuron additive shift. C. A regularization penalty. D. A learning rate.

Select an answer to check.

Answer: Per-neuron additive shift.

Per-neuron additive shift. is the correct answer here. Adds expressiveness. It fits the requirement in the prompt about how is a bias term best characterized. A quick elimination of partially true options helps confirm it.

Q21. Which option best describes forward pass?

A. Backpropagation only. B. Compute outputs from inputs through the network. C. Optimizer step only. D. Gradient computation only.

Select an answer to check.

Answer: Compute outputs from inputs through the network.

Here, Compute outputs from inputs through the network. is the right choice. Used during inference and training. This is the most accurate statement for which option best describes forward pass. A quick elimination of partially true options helps confirm it.

Q22. What is the primary purpose of forward pass?

A. Optimizer step only. B. Gradient computation only. C. Backpropagation only. D. Compute outputs from inputs through the network.

Select an answer to check.

Answer: Compute outputs from inputs through the network.

In this case, Compute outputs from inputs through the network. is correct. Used during inference and training. This is the most accurate statement for what is the primary purpose of forward pass. A quick elimination of partially true options helps confirm it.

Q23. Which statement about forward pass is most accurate?

A. Compute outputs from inputs through the network. B. Backpropagation only. C. Gradient computation only. D. Optimizer step only.

Select an answer to check.

Answer: Compute outputs from inputs through the network.

The best option here is Compute outputs from inputs through the network.. Used during inference and training. This is the most accurate statement for which statement about forward pass is most accurate. A quick elimination of partially true options helps confirm it.

Q24. How is forward pass best characterized?

A. Gradient computation only. B. Optimizer step only. C. Backpropagation only. D. Compute outputs from inputs through the network.

Select an answer to check.

Answer: Compute outputs from inputs through the network.

For this question, Compute outputs from inputs through the network. is correct. Used during inference and training. This is the most accurate statement for how is forward pass best characterized. A quick elimination of partially true options helps confirm it.

Q25. Which option best describes backward pass?

A. Forward pass. B. Loss computation only. C. Compute gradients via backprop. D. Optimizer step.

Select an answer to check.

Answer: Compute gradients via backprop.

Compute gradients via backprop. is the correct answer here. Drives weight updates. This is the most accurate statement for which option best describes backward pass. A quick elimination of partially true options helps confirm it.

Q26. What is the primary purpose of backward pass?

A. Forward pass. B. Optimizer step. C. Loss computation only. D. Compute gradients via backprop.

Select an answer to check.

Answer: Compute gradients via backprop.

Here, Compute gradients via backprop. is the right choice. Drives weight updates. It aligns directly with what the question asks about what is the primary purpose of backward pass. The other options are either incomplete or contextually incorrect.

Q27. Which statement about backward pass is most accurate?

A. Loss computation only. B. Compute gradients via backprop. C. Forward pass. D. Optimizer step.

Select an answer to check.

Answer: Compute gradients via backprop.

In this case, Compute gradients via backprop. is correct. Drives weight updates. It aligns directly with what the question asks about which statement about backward pass is most accurate. The other options are either incomplete or contextually incorrect.

Q28. How is backward pass best characterized?

A. Loss computation only. B. Optimizer step. C. Forward pass. D. Compute gradients via backprop.

Select an answer to check.

Answer: Compute gradients via backprop.

The best option here is Compute gradients via backprop.. Drives weight updates. It aligns directly with what the question asks about how is backward pass best characterized. The other options are either incomplete or contextually incorrect.

Q29. Which option best describes an epoch?

A. A model checkpoint event. B. One pass over the entire training dataset. C. A single sample step. D. A single batch step.

Select an answer to check.

Answer: One pass over the entire training dataset.

For this question, One pass over the entire training dataset. is correct. Multiple epochs typically. It aligns directly with what the question asks about which option best describes an epoch. The other options are either incomplete or contextually incorrect.

Q30. What is the primary purpose of an epoch?

A. A single sample step. B. A single batch step. C. A model checkpoint event. D. One pass over the entire training dataset.

Select an answer to check.

Answer: One pass over the entire training dataset.

One pass over the entire training dataset. is the correct answer here. Multiple epochs typically. It aligns directly with what the question asks about what is the primary purpose of an epoch. The other options are either incomplete or contextually incorrect.

Q31. Which statement about an epoch is most accurate?

A. A model checkpoint event. B. A single sample step. C. One pass over the entire training dataset. D. A single batch step.

Select an answer to check.

Answer: One pass over the entire training dataset.

Here, One pass over the entire training dataset. is the right choice. Multiple epochs typically. This matches the core idea being tested around which statement about an epoch is most accurate. The other options are either incomplete or contextually incorrect.

Q32. How is an epoch best characterized?

A. A model checkpoint event. B. A single sample step. C. One pass over the entire training dataset. D. A single batch step.

Select an answer to check.

Answer: One pass over the entire training dataset.

In this case, One pass over the entire training dataset. is correct. Multiple epochs typically. This matches the core idea being tested around how is an epoch best characterized. The other options are either incomplete or contextually incorrect.

Q33. Which option best describes a mini-batch?

A. One sample per step (SGD). B. A test set sample. C. The whole dataset per step (full batch). D. Small subset of training data per gradient step.

Select an answer to check.

Answer: Small subset of training data per gradient step.

The best option here is Small subset of training data per gradient step.. Common in practice. This matches the core idea being tested around which option best describes a mini-batch. The other options are either incomplete or contextually incorrect.

Q34. What is the primary purpose of a mini-batch?

A. A test set sample. B. Small subset of training data per gradient step. C. One sample per step (SGD). D. The whole dataset per step (full batch).

Select an answer to check.

Answer: Small subset of training data per gradient step.

For this question, Small subset of training data per gradient step. is correct. Common in practice. This matches the core idea being tested around what is the primary purpose of a mini-batch. The other options are either incomplete or contextually incorrect.

Q35. Which statement about a mini-batch is most accurate?

A. Small subset of training data per gradient step. B. The whole dataset per step (full batch). C. A test set sample. D. One sample per step (SGD).

Select an answer to check.

Answer: Small subset of training data per gradient step.

Small subset of training data per gradient step. is the correct answer here. Common in practice. This matches the core idea being tested around which statement about a mini-batch is most accurate. The other options are either incomplete or contextually incorrect.

Q36. How is a mini-batch best characterized?

A. A test set sample. B. Small subset of training data per gradient step. C. The whole dataset per step (full batch). D. One sample per step (SGD).

Select an answer to check.

Answer: Small subset of training data per gradient step.

Here, Small subset of training data per gradient step. is the right choice. Common in practice. That is exactly the concept behind how is a mini-batch best characterized in this context. The other options are either incomplete or contextually incorrect.

Q37. Which option best describes SGD?

A. Newton's method. B. Stochastic gradient descent on mini-batches. C. Full-batch gradient descent only. D. A pure RL update.

Select an answer to check.

Answer: Stochastic gradient descent on mini-batches.

In this case, Stochastic gradient descent on mini-batches. is correct. Foundational optimizer. That is exactly the concept behind which option best describes sgd in this context. The other options are either incomplete or contextually incorrect.

Q38. What is the primary purpose of SGD?

A. Stochastic gradient descent on mini-batches. B. A pure RL update. C. Newton's method. D. Full-batch gradient descent only.

Select an answer to check.

Answer: Stochastic gradient descent on mini-batches.

The best option here is Stochastic gradient descent on mini-batches.. Foundational optimizer. That is exactly the concept behind what is the primary purpose of sgd in this context. The other options are either incomplete or contextually incorrect.

Q39. Which statement about SGD is most accurate?

A. Full-batch gradient descent only. B. Stochastic gradient descent on mini-batches. C. A pure RL update. D. Newton's method.

Select an answer to check.

Answer: Stochastic gradient descent on mini-batches.

For this question, Stochastic gradient descent on mini-batches. is correct. Foundational optimizer. That is exactly the concept behind which statement about sgd is most accurate in this context. The other options are either incomplete or contextually incorrect.

Q40. How is SGD best characterized?

A. A pure RL update. B. Stochastic gradient descent on mini-batches. C. Newton's method. D. Full-batch gradient descent only.

Select an answer to check.

Answer: Stochastic gradient descent on mini-batches.

Stochastic gradient descent on mini-batches. is the correct answer here. Foundational optimizer. That is exactly the concept behind how is sgd best characterized in this context. The other options are either incomplete or contextually incorrect.

Q41. Which option best describes momentum?

A. Adam optimizer entirely. B. Random restarts. C. Adds a fraction of previous update to the current. D. Pure SGD with no memory.

Select an answer to check.

Answer: Adds a fraction of previous update to the current.

Here, Adds a fraction of previous update to the current. is the right choice. Smooths and speeds convergence. It fits the requirement in the prompt about which option best describes momentum. The other options are either incomplete or contextually incorrect.

Q42. What is the primary purpose of momentum?

A. Random restarts. B. Pure SGD with no memory. C. Adds a fraction of previous update to the current. D. Adam optimizer entirely.

Select an answer to check.

Answer: Adds a fraction of previous update to the current.

In this case, Adds a fraction of previous update to the current. is correct. Smooths and speeds convergence. It fits the requirement in the prompt about what is the primary purpose of momentum. The other options are either incomplete or contextually incorrect.

Q43. Which statement about momentum is most accurate?

A. Adam optimizer entirely. B. Random restarts. C. Adds a fraction of previous update to the current. D. Pure SGD with no memory.

Select an answer to check.

Answer: Adds a fraction of previous update to the current.

The best option here is Adds a fraction of previous update to the current.. Smooths and speeds convergence. It fits the requirement in the prompt about which statement about momentum is most accurate. The other options are either incomplete or contextually incorrect.

Q44. How is momentum best characterized?

A. Adam optimizer entirely. B. Pure SGD with no memory. C. Adds a fraction of previous update to the current. D. Random restarts.

Select an answer to check.

Answer: Adds a fraction of previous update to the current.

For this question, Adds a fraction of previous update to the current. is correct. Smooths and speeds convergence. It fits the requirement in the prompt about how is momentum best characterized. The other options are either incomplete or contextually incorrect.

Q45. Which option best describes a loss surface?

A. A confusion matrix. B. A ROC curve. C. A learning curve. D. Landscape of loss over parameter space.

Select an answer to check.

Answer: Landscape of loss over parameter space.

Landscape of loss over parameter space. is the correct answer here. Non-convex for deep nets. It fits the requirement in the prompt about which option best describes a loss surface. The other options are either incomplete or contextually incorrect.

Q46. What is the primary purpose of a loss surface?

A. A confusion matrix. B. A ROC curve. C. A learning curve. D. Landscape of loss over parameter space.

Select an answer to check.

Answer: Landscape of loss over parameter space.

Here, Landscape of loss over parameter space. is the right choice. Non-convex for deep nets. This is the most accurate statement for what is the primary purpose of a loss. The other options are either incomplete or contextually incorrect.

Q47. Which statement about a loss surface is most accurate?

A. A confusion matrix. B. Landscape of loss over parameter space. C. A ROC curve. D. A learning curve.

Select an answer to check.

Answer: Landscape of loss over parameter space.

In this case, Landscape of loss over parameter space. is correct. Non-convex for deep nets. This is the most accurate statement for which statement about a loss surface is most. The other options are either incomplete or contextually incorrect.

Q48. How is a loss surface best characterized?

A. A confusion matrix. B. A learning curve. C. Landscape of loss over parameter space. D. A ROC curve.

Select an answer to check.

Answer: Landscape of loss over parameter space.

The best option here is Landscape of loss over parameter space.. Non-convex for deep nets. This is the most accurate statement for how is a loss surface best characterized. The other options are either incomplete or contextually incorrect.

Q49. Which option best describes a local minimum?

A. The global minimum. B. Point lower than its neighborhood but not globally. C. An inflection point. D. A saddle point.

Select an answer to check.

Answer: Point lower than its neighborhood but not globally.

For this question, Point lower than its neighborhood but not globally. is correct. Can trap optimizers. This is the most accurate statement for which option best describes a local minimum. The other options are either incomplete or contextually incorrect.

Q50. What is the primary purpose of a local minimum?

A. An inflection point. B. A saddle point. C. The global minimum. D. Point lower than its neighborhood but not globally.

Select an answer to check.

Answer: Point lower than its neighborhood but not globally.

Point lower than its neighborhood but not globally. is the correct answer here. Can trap optimizers. This is the most accurate statement for what is the primary purpose of a local. The other options are either incomplete or contextually incorrect.