Tweeted By @OriolVinyalsML

Rapid Learning or Feature Reuse? Meta-learning algorithms on standard benchmarks have much more feature reuse than rapid learning! This also gives us a way to simplify MAML -- (Almost) No Inner Loop (A)NIL. https://t.co/aVFornxFJ3 With Aniruddh Raghu @maithra_raghu Samy Bengio. pic.twitter.com/7T6SzMYfiY

— Oriol Vinyals (@OriolVinyalsML) September 23, 2019