Computer — All MCQs
4795 questions available
Q1961.Which AI subfield enables machines to make decisions by applying rules like human experts?
Correct Answer: D
Expert Systems represent a classic AI field that mimics the logical decision-making of human domain experts using predefined rule bases.
Q1962.An expert system mainly consists of a knowledge base and which of the following?
Correct Answer: B
The inference engine is the brain of an expert system, applying logical rules to the knowledge base to deduce solutions.
Q1963.Which type of neural network is best suited for sequential, temporal, or time-series data?
Correct Answer: A
Recurrent Neural Networks (RNNs) feature internal memory loops, making them ideal for processing sequence data like text and time-series.
Q1964.What is the vanishing gradient problem associated with?
Correct Answer: A
Vanishing gradients occur when error signals shrink exponentially as they propagate backward through deep networks, stopping early layers from learning.
Q1965.Which loss function is most commonly used for classification problems?
Correct Answer: D
Cross-Entropy loss (log loss) measures the performance of a classification model whose output is a probability value between 0 and 1.
Q1966.Which company developed the LLaMA family of large language models?
Correct Answer: C
LLaMA (Large Language Model Meta AI) is an open-access foundational model family developed and released by Meta.
Q1967.What is 'tokenization' in NLP?
Correct Answer: A
Tokenization is the process of splitting raw text strings into smaller meaningful units like words or subwords before feeding them into an NLP model.
Q1968.Which technique generates new realistic data by setting two neural networks in competition?
Correct Answer: A
Generative Adversarial Networks (GANs) utilize a generator (creating fakes) and a discriminator (checking authenticity) in a game-theoretic competition.
Q1969.In a GAN, the two main competing components are the generator and the?
Correct Answer: B
The discriminator acts as a classifier, trying to distinguish between real training data and synthetic data produced by the generator.
Q1970.What is 'transfer learning' in machine learning?
Correct Answer: C
Transfer learning leverages knowledge learned from training on a massive dataset to accelerate learning on a new, related task.
Q1971.Which evaluation metric measures the average squared difference between predictions and actual values in regression?
Correct Answer: B
Mean Squared Error (MSE) is a standard metric used to measure model accuracy for continuous regression tasks.
Q1972.Which of the following best describes 'reinforcement learning'?
Correct Answer: C
Reinforcement learning involves an agent interacting with an environment, learning to make decisions by maximizing cumulative rewards.
Q1973.What does the term 'feature' mean in machine learning?
Correct Answer: C
A feature is an explanatory variable or individual input column used by machine learning models to make predictions.
Q1974.Which AI large language model family powers Anthropic's chatbot?
Correct Answer: B
Claude is Anthropic's flagship family of highly advanced large language models.
Q1975.What is 'data augmentation' commonly used for?
Correct Answer: D
Data augmentation applies random transformations (like rotation, scaling, or cropping) to existing training data to create diverse new samples.
Q1976.Which machine learning algorithm is a supervised classification method based on finding optimal separating hyperplanes?
Correct Answer: A
Support Vector Machine (SVM) is a supervised algorithm that searches for the optimal hyperplane maximizing the margins between classes.
Q1977.What is the role of the 'learning rate' in training deep neural networks?
Correct Answer: D
The learning rate is a critical hyperparameter that dictates the step size taken by the optimization algorithm (like SGD) toward a minimum loss.
Q1978.What is 'sentiment analysis' in NLP?
Correct Answer: A
Sentiment analysis uses NLP to identify, extract, and study subjective emotional tones within text data.
Q1979.Which method partitions data to evaluate model performance and generalization reliably?
Correct Answer: B
Cross-validation is a statistical resampling technique that splits data into folds to train and test models iteratively, ensuring robust performance evaluation.
Q1980.In digital forensics, AI is primarily used to?
Correct Answer: B
AI systems assist forensic investigators by quickly searching, categorizing, and filtering through massive sets of hard drive or chat log evidence.