Neurips 2019 Reproducibility Challenge

  • Course project for 2019 Fall COMP 551 Applied Machine Learning
  • We conducted extensive ablation study of this paper to examine its claims’ soundness and its model’s robustness
  • We identified certain drawbacks that undermine the reproducibility of this paper
  • Our report, despite being rejected for final journal publication, is well appreciated by reviewers
  • Project report availble here
Avatar
Bruce (Zhi) Wen
Master Student in Computer Science

Publications

In our ablation study, we reported several issues regarding the model implementation, and examined the effect of latent space dimension and number of Transformer layers on the performance of the original model.
In NeurIPS 2019 Reproducibility Challenge, 2019