Technological Knowledge MCQs with Answers and Explanations

Which of these models is specifically designed for image recognition tasks?

A. RNN
B. CNN
C. LSTM
D. Naive Bayes

Explanation:
Convolutional Neural Networks (CNNs) are designed for processing image data.

What is 'word embedding' in natural language processing?

A. A way to vectorize words, capturing semantics
B. A web embedding in HTML
C. A type of encryption
D. A grammar technique

Explanation:
Word embeddings convert words into numeric vectors that capture semantic relationships.

Which of the following is a common loss function for regression problems?

A. Cross-entropy loss
B. Mean squared error
C. Hinge loss
D. Binary entropy

Explanation:
Mean squared error is typically used for regression to measure the average squared difference.

What is the main goal of 'dimensionality reduction' techniques like PCA?

A. Increase the number of features
B. Decrease model accuracy
C. Remove irrelevant or correlated features
D. Encrypt data

Explanation:
Dimensionality reduction techniques reduce feature count by eliminating redundancy or irrelevance.

Which algorithm is best suited for classification tasks in machine learning?

A. Linear regression
B. K-means
C. Support vector machine
D. PCA

Explanation:
Support Vector Machine is commonly used for classification problems.

In neural networks, what is 'backpropagation' used for?

A. Adjusting weights using gradient descent
B. Propagating inputs through layers
C. Generating output predictions
D. Normalizing data

Explanation:
Backpropagation computes the gradients of the loss with respect to the network weights, which are then used to update them via optimization methods.

What is 'reinforcement learning' primarily used for?

A. Classifying data into labels
B. Reducing dimensionality
C. Learning to make decisions via rewards
D. Clustering unlabeled data

Explanation:
Reinforcement learning is about learning optimal actions based on rewards and penalties from an environment.

What is the purpose of a 'validation set' during model training?

A. To train the model parameters
B. To evaluate the final model performance
C. To tune hyperparameters and avoid overfitting
D. To serve as training data

Explanation:
The validation set is used to tune hyperparameters and detect overfitting during training.

In machine learning, what is 'dropout'?

A. Dropping the dataset
B. Randomly omitting neurons during training
C. Pruning decision trees
D. Deleting old data

Explanation:
Dropout is a regularization technique in which random neurons are randomly deactivated during training to prevent overfitting.

Which of the following is a transformer-based language model developed by Google?

A. GPT
B. BERT
C. ResNet
D. AlexNet

Explanation:
BERT is a transformer-based model by Google for NLP tasks.