2 Matching Annotations
  1. Oct 2024
  2. mlpr.inf.ed.ac.uk mlpr.inf.ed.ac.uk
    MLPR w4b - Machine Learning and Pattern Recognition
    1
    1. DavidKrame 10 Oct 2024
      it can be shown that the posterior mean corresponds to using an L2 regularized model

      There are other types of regularization that work in practice (for example, L1, which is not even differentiable). Does this mean that L2 regularization is the best ?

    Visit annotations in context

    Annotators

    URL

    mlpr.inf.ed.ac.uk/2024/notes/w4b_bayesian_inference_prediction.html
  3. mlpr.inf.ed.ac.uk mlpr.inf.ed.ac.uk
    MLPR w4a - Machine Learning and Pattern Recognition
    1
    1. DavidKrame 08 Oct 2024
      specifying which models we think are plausible before observing any data

      Does this corresponds to what we usually call initialisation ?

    Visit annotations in context

    Annotators

    URL

    mlpr.inf.ed.ac.uk/2024/notes/w4a_bayesian_regression.html