[Re] Learning to Learn By Self-Critique

Isac Arnekvist, Dmytro Kalpakchi

Abstract: This work is a reproducibility study of the paper of Antoniou and Storkey [2019], published at NeurIPS 2019. Our results are in parts similar to the ones reported in the original paper, supporting the central claim of the paper that the proposed novel method, called Self-Critique and Adapt (SCA), improves the performance of MAML++. The conducted additional experiments on the Caltech-UCSD Birds 200 dataset confirm the superiority of SCA compared to MAML++. The ability to reproduce these results, without many of the implementation details, we believe is in favor of the proposed method. In addition, the reproduced paper suggests a novel high-end version of MAML++ for which we could not reproduce the same results. Our conclusion is that the reproduced paper is not self-contained, and we identify this as a potential reason of failure to reproduce all results.

Track: Replicability

NeurIPS Paper Id: https://openreview.net/forum?id=HJfHJBSlLS