Texas Descent And Distribution

Model-Agnostic Meta-Learning for Fast Adaptation of …

using one or more gradient descent updates on task T i. For example, when using one gradient update, 0 i = r L T i (f): The step size may be fixed as a hyperparameter or meta-learned. For simplicity of notation, we will consider one gradient update for the rest of this section, but using multi-ple gradient updates is a straightforward extension.

  Model, Team, Learning, Fast, Adaptation, Descent, Agnostics, Model agnostic meta learning for fast adaptation