Selecmix: Debiased learning by contradicting-pair sampling

I Hwang, S Lee, Y Kwak, SJ Oh… - Advances in …, 2022 - proceedings.neurips.cc
Neural networks trained with ERM (empirical risk minimization) sometimes learn unintended
decision rules, in particular when their training data is biased, ie, when training labels are …
被引用次数：18 相关文章

SelecMix: debiased learning by contradicting-pair sampling

I Hwang, S Lee, Y Kwak, SJ Oh, D Teney… - Proceedings of the 36th …, 2022 - dl.acm.org
Neural networks trained with ERM (empirical risk minimization) sometimes learn unintended
decision rules, in particular when their training data is biased, ie, when training labels are …

SelecMix: Debiased Learning by Contradicting-pair Sampling

I Hwang, S Lee, Y Kwak, SJ Oh, D Teney… - Advances in Neural …, 2022 - openreview.net
Neural networks trained with ERM (empirical risk minimization) sometimes learn unintended
decision rules, in particular when their training data is biased, ie, when training labels are …

SelecMix: Debiased Learning by Contradicting-pair Sampling

I Hwang, S Lee, Y Kwak, SJ Oh, D Teney… - arXiv e …, 2022 - ui.adsabs.harvard.edu
Neural networks trained with ERM (empirical risk minimization) sometimes learn unintended
decision rules, in particular when their training data is biased, ie, when training labels are …

SelecMix: Debiased Learning by Contradicting-pair Sampling

I Hwang, S Lee, Y Kwak, SJ Oh, D Teney… - arXiv preprint arXiv …, 2022 - arxiv.org
Neural networks trained with ERM (empirical risk minimization) sometimes learn unintended
decision rules, in particular when their training data is biased, ie, when training labels are …