Week 6: Intro to Machine Learning Flashcards
What is machine learning (ML)?
A process where systems improve performance on a task through experience.
Give Tom Mitchell’s formal definition of learning.
A computer program learns from experience E with respect to task T and performance measure P if its performance on T, as measured by P, improves with experience E.
Why is ML important in modern AI?
Because many tasks are too complex to program by hand and environments can change over time.
List the four main types of ML.
Supervised learning, unsupervised learning, reinforcement learning, semi-supervised learning.
What is supervised learning?
Learning from labeled input-output pairs to infer a function mapping inputs to outputs.
What is unsupervised learning?
Learning patterns or structure from input data without labeled outputs (e.g., clustering).
What is reinforcement learning?
Learning by receiving feedback in the form of rewards or penalties after actions are taken.
What is semi-supervised learning?
Learning from a small number of labeled examples combined with many unlabeled examples.
What is inductive learning?
Learning a general function from specific examples (input-output pairs).
What is a hypothesis in ML?
A proposed function that maps inputs to outputs, approximating the true function.
What is overfitting?
When a model fits the training data too well but fails to generalize to new data.
What is the hypothesis space?
The set of all possible functions or models that can be chosen during learning.
What is generalization in ML?
A model’s ability to perform well on new, unseen examples.
What is a decision tree?
A tree-structured model that makes decisions by testing attributes and assigning outputs.
What is entropy in decision tree learning?
A measure of uncertainty or impurity in a set of examples.
What is information gain?
The reduction in entropy achieved by splitting on a particular attribute.
What is the purpose of pruning in decision trees?
To reduce overfitting by removing nodes that add little predictive power.
What is a consistent hypothesis?
A hypothesis that correctly predicts the output for all training examples.
What is the goal of supervised learning?
To learn a function h(x) that closely approximates the true function f(x).
What is regression in ML?
A type of supervised learning where the output is a continuous value.
What is classification in ML?
A type of supervised learning where the output is a category or class label.
What is linear regression?
A model that predicts a continuous output as a linear function of input features.
What is the cost function in linear regression?
A function (e.g., Mean Squared Error) used to measure prediction error and guide learning.
What is gradient descent?
An optimization algorithm used to minimize the cost function by updating parameters iteratively.
