📚 node [[convex_function|convex function]]
Welcome! Nobody has contributed anything to 'convex_function|convex function' yet. You can:
  • Write something in the document below!
    • There is at least one public document in every node in the Agora. Whatever you write in it will be integrated and made available for the next visitor to read and edit.
  • Write to the Agora from social media.
    • If you follow Agora bot on a supported platform and include the wikilink [[convex_function|convex function]] in a post, the Agora will link it here and optionally integrate your writing.
  • Sign up as a full Agora user.
    • As a full user you will be able to contribute your personal notes and resources directly to this knowledge commons. Some setup required :)
⥅ related node [[convex_function]]
⥅ node [[convex_function]] pulled by Agora

convex function

Go back to the [[AI Glossary]]

A function in which the region above the graph of the function is a convex set. The prototypical convex function is shaped something like the letter U. For example, the following are all convex functions:

Graphical representations of convex functions

A typical convex function is shaped like the letter 'U'.

By contrast, the following function is not convex. Notice how the region above the graph is not a convex set:

A graphical representation of a non-convex function

A strictly convex function has exactly one local minimum point, which is also the global minimum point. The classic U-shaped functions are strictly convex functions. However, some convex functions (for example, straight lines) are not U-shaped.

A lot of the common loss functions, including the following, are convex functions:

  • L2 loss
  • Log Loss
  • L1 regularization
  • L2 regularization

Many variations of gradient descent are guaranteed to find a point close to the minimum of a strictly convex function. Similarly, many variations of stochastic gradient descent have a high probability (though, not a guarantee) of finding a point close to the minimum of a strictly convex function.

The sum of two convex functions (for example, L2 loss + L1 regularization) is a convex function.

Deep models are never convex functions. Remarkably, algorithms designed for convex optimization tend to find reasonably good solutions on deep networks anyway, even though those solutions are not guaranteed to be a global minimum.

📖 stoas
⥱ context