1 5
1 5
Answer: AI is the simulation of human intelligence in machines programmed to think and learn. It
enables machines to perform tasks typically requiring human intelligence, like visual perception,
speech recognition, and decision-making.
Answer: Neural networks are computational models inspired by the human brain. They consist of
interconnected nodes (neurons) that process data in layers, enabling deep learning and other
complex AI functions.
Answer: In supervised learning, the model is trained on labeled data, where each input has a known
output. In unsupervised learning, the model is trained on data without labeled responses, aiming to
find patterns or groupings.
Answer: Activation functions determine whether a neuron should be activated. They introduce non-
linear properties to the network, enabling it to learn complex data patterns.
Answer: It’s a type of ML where an agent learns by interacting with its environment, making
decisions, and receiving feedback (rewards or penalties) to maximize a goal.
Answer: Transfer learning involves reusing a pre-trained model on a new, similar problem, saving
time and resources, especially when data for the new problem is limited.
Answer: Overfitting occurs when a model learns the training data too well, including noise, and
performs poorly on new data. Underfitting happens when a model is too simple, failing to capture
the data’s patterns, resulting in poor performance.
Answer: Regularization involves adding a penalty term to the loss function to prevent overfitting.
Common techniques include L1 (Lasso) and L2 (Ridge) regularization.
11. 11. Describe a decision tree.
Answer: A decision tree is a model that splits data into branches based on feature values, leading to
decision nodes, and ultimately a prediction outcome. It’s easy to interpret but can be prone to
overfitting.
Answer: Hyperparameters are configuration settings external to the model that need to be set
before training (like learning rate and number of epochs). They differ from parameters learned by
the model.
Answer: Gradient descent is an optimization algorithm that minimizes the cost function by
iteratively adjusting parameters in the direction that reduces error.
Answer: Backpropagation is an algorithm used to train neural networks. It calculates the gradient of
the error with respect to weights, adjusting them to minimize the error.
Answer: CNNs are specialized neural networks for image data, using convolutional layers to
automatically detect patterns in images.
Answer: Proposed by Alan Turing, the Turing Test evaluates if a machine's behavior is
indistinguishable from a human’s.
Answer: AI Winter refers to periods when AI research and funding stagnated due to unmet
expectations.
Answer: It’s a computer system that mimics human expertise in specific fields, using a knowledge
base and inference rules.
Answer: Milestones include the creation of neural networks, the Turing Test, IBM’s Deep Blue beating a
chess grandmaster, and recent advances in deep learning.
21. What is logistic regression?
Answer: Logistic regression is a classification algorithm used to model binary outcomes, predicting
probabilities between 0 and 1.
Answer: k-NN is a non-parametric algorithm that classifies data points based on the closest training
examples in the feature space.
Answer: SVM is a classification algorithm that finds the hyperplane that best separates classes in
high-dimensional space.
Answer: Naïve Bayes is a probabilistic classifier based on Bayes’ theorem, assuming independence
between features.
Answer: K-Means is an unsupervised algorithm that partitions data into k clusters, with each point
assigned to the nearest cluster center.
Answer: Random Forest is an ensemble learning method that builds multiple decision trees and
averages their predictions for improved accuracy.
Answer: PCA is a dimensionality reduction technique that transforms data into principal
components, reducing feature space.
Answer: Ensemble methods, like boosting and bagging, combine multiple algorithms to improve
predictive performance and robustness.
Answer: GBM is an ensemble method that builds models sequentially, focusing on correcting errors
of previous models to improve accuracy.
Answer: XGBoost is an optimized gradient boosting algorithm for efficiency, accuracy, and speed,
popular in machine learning competitions.
Answer: Transformers are neural network architectures based on self-attention mechanisms, widely
used in natural language processing tasks.
46. What is Mean Absolute Error (MAE) and Mean Squared Error (MSE)?
Answer: MAE measures average absolute errors, while MSE measures the average squared
differences, with MSE penalizing larger errors more heavily.
Answer: The learning rate determines the step size in gradient descent, affecting model
convergence speed and accuracy.
Answer: The bias-variance tradeoff is the balance between a model’s accuracy on training data (low
bias) and its ability to generalize to new data (low variance).