A GPFlow implementation of Meta-learning GPs (Fortuin & R\“atsch, 2019).
NP: during training, sees { xy_1, …, xyn+m } The number of contexts (n) and number of targets (m) are chosen randomly at each iteration (n ∼ U [3, 100], m ∼ n + U [0, 100 − n]). Each x-value is drawn uniformly at random in [−2, 2].
Bibliography
Fortuin, V., & R\“atsch, Gunnar (2019), Deep mean functions for meta-learning in gaussian processes, CoRR. ↩