Machine Learning Introduction to Machine Learning Syllabus

Introduction to Machine Learning syllabus — full chapter and topic list

Expand any chapter to see its topics and sub-topics. This is the whole examinable outline for Introduction to Machine Learning in Machine Learning, not a summary of it.

1. 1

   Overview of Machine Learning

   4 topics

   • Definition of Machine Learning
   • Types of Machine Learning
   • Applications of Machine Learning
   • History of Machine Learning
2. 2

   Basic Concepts

   4 topics

   • Supervised Learning
   • Unsupervised Learning
   • Reinforcement Learning
   • Semi-Supervised Learning

Introduction to Machine Learning flashcards for Machine Learning

24 of 49 cards from the Introduction to Machine Learning deck — real questions with worked answers.

1. What is the standard definition of Machine Learning?

   Machine Learning is a subfield of artificial intelligence in which systems automatically learn patterns from data and improve their performance on a task through experience, without being explicitly programmed for that task.

2. What is Tom Mitchell's formal (well-posed) definition of a learning problem?

   A computer program learns from experience $E$ with respect to some class of tasks $T$ and performance measure $P$, if its performance at tasks in $T$, as measured by $P$, improves with experience $E$.

3. How does Machine Learning differ from traditional rule-based programming?

   In traditional programming a human writes explicit rules: data plus program produce output. In machine learning the system is given data and the desired outputs and infers the rules (the model) itself: data plus output produce the program.

4. What are the three main types (paradigms) of Machine Learning?

   Supervised learning, unsupervised learning, and reinforcement learning. (Semi-supervised learning is often listed as a fourth, hybrid paradigm.)

5. What distinguishes the learning paradigms by the kind of feedback they use?

   Supervised learning uses labeled feedback (correct answers), unsupervised learning uses no labels (only input structure), and reinforcement learning uses evaluative feedback in the form of rewards or penalties from interacting with an environment.

6. In Machine Learning terminology, what is a 'feature' and what is a 'label'?

   A feature is an individual measurable input attribute of an example (a column / independent variable $x$). A label is the target output value $y$ that a supervised model is trained to predict.

7. List five major application areas of Machine Learning.

   Examples include: image and facial recognition (computer vision), speech recognition and natural language processing, recommendation systems, spam/fraud detection, medical diagnosis, autonomous vehicles, and financial forecasting.

8. Which ML paradigm typically powers product recommendation systems and how?

   Recommendation systems mainly use unsupervised/collaborative-filtering techniques (finding patterns and similarities among users and items) and sometimes supervised learning to predict ratings or click probabilities.

9. Who coined the term 'Machine Learning' and in what year?

   Arthur Samuel coined the term 'Machine Learning' in 1959, based on his work on a self-learning checkers (draughts) playing program.

10. What 1950 contribution by Alan Turing is foundational to machine learning and AI?

    Alan Turing proposed the 'Imitation Game' (the Turing Test) in his 1950 paper 'Computing Machinery and Intelligence', and raised the idea of machines that can learn.

11. What was the Perceptron, who created it, and in what year?

    The Perceptron was an early single-layer neural network model for binary classification, invented by Frank Rosenblatt in 1957 (introduced 1958).

12. What event caused the 'AI winter' relevant to early neural network research?

    Minsky and Papert's 1969 book 'Perceptrons' showed that single-layer perceptrons could not solve non-linearly-separable problems such as XOR, which reduced funding and interest in neural networks for years.

13. What algorithmic breakthrough in the 1980s revived neural network research?

    The backpropagation algorithm for training multilayer neural networks, popularized by Rumelhart, Hinton, and Williams in 1986, enabled efficient learning in deep networks.

14. What is supervised learning?

    Supervised learning trains a model on a dataset of input-output pairs (labeled examples) so it can learn a mapping $f: X \to Y$ that predicts the correct output $y$ for new, unseen inputs $x$.

15. What are the two main task types in supervised learning?

    Classification (predicting a discrete categorical label) and regression (predicting a continuous numerical value).

16. Give the difference between classification and regression with examples.

    Classification predicts discrete categories (e.g., spam vs. not spam, cat vs. dog). Regression predicts continuous quantities (e.g., house price, temperature). Classification outputs labels; regression outputs real numbers.

17. Name four common supervised learning algorithms.

    Linear regression, logistic regression, decision trees, support vector machines (SVM), k-nearest neighbors (KNN), naive Bayes, and random forests are all common supervised algorithms.

18. Write the hypothesis (model) equation for simple linear regression.

    $$\hat{y} = \theta_0 + \theta_1 x$$ where $\theta_0$ is the intercept (bias) and $\theta_1$ is the slope (weight).

19. What is the Mean Squared Error (MSE) cost function used in regression?

    $$J(\theta) = \frac{1}{m} \sum_{i=1}^{m} \left( \hat{y}^{(i)} - y^{(i)} \right)^{2}$$ where $m$ is the number of training examples, $\hat{y}^{(i)}$ is the prediction, and $y^{(i)}$ is the true value.

20. What is the sigmoid (logistic) function used in logistic regression, and its output range?

    $$\sigma(z) = \frac{1}{1 + e^{-z}}$$ It maps any real input to the range $(0, 1)$, making it suitable for outputting probabilities.

21. State the gradient descent parameter update rule.

    $$\theta_j := \theta_j - \alpha \frac{\partial}{\partial \theta_j} J(\theta)$$ where $\alpha$ is the learning rate controlling the step size.

22. What role does the learning rate $\alpha$ play in gradient descent?

    It controls the step size of each update. If $\alpha$ is too large, the algorithm may overshoot or diverge; if too small, convergence is very slow.

23. What is unsupervised learning?

    Unsupervised learning finds hidden patterns, structure, or groupings in input data that has no labeled outputs, discovering the inherent organization of the data on its own.

24. What are the main task types in unsupervised learning?

    Clustering (grouping similar data points), dimensionality reduction (compressing features while preserving structure), and association rule mining (finding relationships among variables).

See more Introduction to Machine Learning flashcards →

Planning Introduction to Machine Learning for Machine Learning

Introduction to Machine Learning (Machine Learning) FAQ