Skip to content

Instantly share code, notes, and snippets.

@veekaybee
Created February 1, 2023 17:16
Show Gist options
  • Save veekaybee/ace6d62da2e72ce424b78dc730a47ce9 to your computer and use it in GitHub Desktop.
Save veekaybee/ace6d62da2e72ce424b78dc730a47ce9 to your computer and use it in GitHub Desktop.

Screen Shot 2023-02-01 at 12 05 27 PM

Algorithms find the best ways to do things, but they don't explain "how" they came to those conclusions.

Screen Shot 2023-02-01 at 12 07 02 PM

This is a common way to formulate ML problems, using target functions that we don't know but we want to approximate and learn.

Screen Shot 2023-02-01 at 12 08 27 PM

Screen Shot 2023-02-01 at 12 09 53 PM

The choice of algorithm (e.g. neural network) and the configuration of the algorithm (e.g. network topology and hyperparameters) define the space of possible hypothesis that the model may represent.

Learning is a search through the space of possible hypotheses for one that will perform well, even on new examples beyond the training set.


    h (hypothesis): A single hypothesis, e.g. an instance or specific candidate model that maps inputs to outputs and can be evaluated and used to make predictions.
    H (hypothesis set): A space of possible hypotheses for mapping inputs to outputs that can be searched, often constrained by the choice of the framing of the problem, the choice of model and the choice of model configuration.

Screen Shot 2023-02-01 at 12 13 09 PM

Supervised, Unsupervised and RL

Screen Shot 2023-02-01 at 12 13 57 PM

Screen Shot 2023-02-01 at 12 14 21 PM

Screen Shot 2023-02-01 at 12 14 37 PM

Hoeffding's Inequality

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment