A First Course in Machine Learning, 2nd
Edition (2015) – Girolami – Solutions
Manual PDF
What is machine learning? - answer-Machine learning is the task of learning a function f(x) that
maps input features to outputs by fitting patterns in data.
What does it mean to "learn a curve" in ML? - answer-It means learning the relationship
between input features and output values so the model can make predictions on unseen data.
What is supervised learning? - answer-A type of ML where training data includes labels y. The
model learns to map inputs x to outputs y.
What is unsupervised learning? - answer-A type of ML where there are no labels. The goal is to
discover structure or patterns in the data.
What is regression? - answer-A supervised learning task where the output is a continuous real-
valued number.
What is classification? - answer-A supervised learning task where the output is a discrete class
label.
What is a loss function? - answer-A function that measures how far a model's predictions are
from the true labels.
What is the purpose of a loss function? - answer-It provides a quantitative objective that the
1
, model minimizes during training.
What does a model f(x) represent? - answer-A function that maps input features to predicted
outputs.
What is conditional probability? - answer-The probability of event A occurring given that event
B has already occurred.
What is Bayes' Theorem? - answer-P(A|B) = P(B|A)P(A)/P(B)
What is a prior in Bayes' Theorem? - answer-The belief about an event before seeing the data.
What is likelihood in Bayes' Theorem? - answer-The probability of observing the data given a
hypothesis.
What is the posterior? - answer-The updated belief about an event after observing data.
What is expectation? - answer-The average (mean) value of a random variable.
What is variance? - answer-A measure of how spread out values are around the mean.
What is covariance? - answer-A measure of how two random variables move together.
What is correlation? - answer-A normalized version of covariance that lies between -1 and 1.
What is the dot product of two vectors? - answer-The sum of element-wise products; measures
2
Edition (2015) – Girolami – Solutions
Manual PDF
What is machine learning? - answer-Machine learning is the task of learning a function f(x) that
maps input features to outputs by fitting patterns in data.
What does it mean to "learn a curve" in ML? - answer-It means learning the relationship
between input features and output values so the model can make predictions on unseen data.
What is supervised learning? - answer-A type of ML where training data includes labels y. The
model learns to map inputs x to outputs y.
What is unsupervised learning? - answer-A type of ML where there are no labels. The goal is to
discover structure or patterns in the data.
What is regression? - answer-A supervised learning task where the output is a continuous real-
valued number.
What is classification? - answer-A supervised learning task where the output is a discrete class
label.
What is a loss function? - answer-A function that measures how far a model's predictions are
from the true labels.
What is the purpose of a loss function? - answer-It provides a quantitative objective that the
1
, model minimizes during training.
What does a model f(x) represent? - answer-A function that maps input features to predicted
outputs.
What is conditional probability? - answer-The probability of event A occurring given that event
B has already occurred.
What is Bayes' Theorem? - answer-P(A|B) = P(B|A)P(A)/P(B)
What is a prior in Bayes' Theorem? - answer-The belief about an event before seeing the data.
What is likelihood in Bayes' Theorem? - answer-The probability of observing the data given a
hypothesis.
What is the posterior? - answer-The updated belief about an event after observing data.
What is expectation? - answer-The average (mean) value of a random variable.
What is variance? - answer-A measure of how spread out values are around the mean.
What is covariance? - answer-A measure of how two random variables move together.
What is correlation? - answer-A normalized version of covariance that lies between -1 and 1.
What is the dot product of two vectors? - answer-The sum of element-wise products; measures
2
