Course

Machine Learning

California Institute of Technology

This is an introductory course by Caltech Professor Yaser Abu-Mostafa on machine learning that covers the basic theory, algorithms, and applications. Machine learning (ML) enables computational systems to adaptively improve their performance with experience accumulated from the observed data. ML techniques are widely applied in engineering, science, finance, and commerce to build systems for which we do not have full mathematical specification (and that covers a lot of systems). The course balances theory and practice, and covers the mathematical as well as the heuristic aspects.

Course Lectures
  • The Learning Problem
    Yaser Abu-Mostafa

    The Learning Problem - Introduction; supervised, unsupervised, and reinforcement learning. Components of the learning problem.

  • Error and Noise
    Yaser Abu-Mostafa

    Error and Noise - The principled choice of error measures. What happens when the target we want to learn is noisy.

  • Training Versus Testing
    Yaser Abu-Mostafa

    Training versus Testing - The difference between training and testing in mathematical terms. What makes a learning model able to generalize?

  • Theory of Generalization
    Yaser Abu-Mostafa

    Theory of Generalization - How an infinite model can learn from a finite sample. The most important theoretical result in machine learning.

  • The VC Dimension
    Yaser Abu-Mostafa

    The VC Dimension - A measure of what it takes a model to learn. Relationship to the number of parameters and degrees of freedom.

  • Bias-Variance Tradeoff
    Yaser Abu-Mostafa

    Bias-Variance Tradeoff - Breaking down the learning performance into competing quantities. The learning curves.

  • The Linear Model II
    Yaser Abu-Mostafa

    The Linear Model II - More about linear models. Logistic regression, maximum likelihood, and gradient descent.

  • Neural Networks
    Yaser Abu-Mostafa

    Neural Networks - A biologically inspired model. The efficient backpropagation learning algorithm. Hidden layers.

  • Overfitting
    Yaser Abu-Mostafa

    Overfitting - Fitting the data too well; fitting the noise. Deterministic noise versus stochastic noise.

  • Regularization
    Yaser Abu-Mostafa

    Regularization - Putting the brakes on fitting the noise. Hard and soft constraints. Augmented error and weight decay.

  • Validation
    Yaser Abu-Mostafa

    Validation - Taking a peek out of sample. Model selection and data contamination. Cross validation.

  • Support Vector Machines
    Yaser Abu-Mostafa

    Support Vector Machines - One of the most successful learning algorithms; getting a complex model at the price of a simple one.

  • Kernel Methods
    Yaser Abu-Mostafa

    Kernel Methods - Extending SVM to infinite-dimensional spaces using the kernel trick, and to non-separable data using soft margins.

  • Radial Basis Functions
    Yaser Abu-Mostafa

    Radial Basis Functions - An important learning model that connects several machine learning models and techniques.

  • Three Learning Principles
    Yaser Abu-Mostafa

    Three Learning Principles - Major pitfalls for machine learning practitioners; Occam's razor, sampling bias, and data snooping.

  • Epilogue
    Yaser Abu-Mostafa

    Epilogue - The map of machine learning. Brief views of Bayesian learning and aggregation methods.