Require: p(T) : distribution over tasks Require: α, β : step size hyperparameters 1: randomly initialize θ 2: while not done do 3: Sample batch of tasks $T_i~p\left(T\right)$ 4: for all do 5: Evaluate ${\nabla }_{\theta }{L}_{T_i}\left(f_{\theta }\right)$ with respect to K examples 6: Compute adapted parameters with gradient descent: $\theta {\text{'}}_i=\theta -a{\nabla }_{\theta }{L}_{T_i}\left(f_{\theta }\right)$ 7: end for 8: Update $\theta \leftarrow \theta -\beta {\nabla }_{\theta }{\text{Σ}}_{T_i~p\left(T\right)}{L}_{T_i}\left(f_{\theta {\text{'}}_i}\right)$ 9: end while