Leakage and the Reproducibility Crisis in ML-based Science — HN submission

Thu, 18 Aug 2022 10:13:25 GMT

Leakage and the Reproducibility Crisis in ML-Based Science

1 point; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.

Leakage and the Reproducibility Crisis in ML-based Science — mentioned by YeGoblynQueenne

Sun, 17 Jul 2022 11:13:16 GMT

I'm reading the draft paper linked from the article and "data leakage" is used multiple times without any attempt at defining it. Oh well, I guess this is only meant for insiders who understand the in-group jargon. The rest of us need not be interested at all. Edit: for context, here is the paragraph titled "Leakage" from the draft paper: Leakage. Data leakage has long been recognized as a lead- ing cause of errors in ML applications (Nisbet et al., 2009). In formative work on leakage, Kaufman et al. (2012) provide an overview of different types of errors and give several rec- ommendations…

Paper and discussion history. A mention is not an endorsement.

Leakage and the Reproducibility Crisis in ML-based Science · HackerPapers discussions

Leakage and the Reproducibility Crisis in ML-based Science — HN submission

Leakage and the Reproducibility Crisis in ML-based Science — mentioned by YeGoblynQueenne