×

MixMatch

swMATH ID: 41280
Software Authors: David Berthelot, Nicholas Carlini, Ian Goodfellow, Nicolas Papernot, Avital Oliver, Colin Raffel
Description: Mixmatch: A holistic approach to semi-supervised learning. Semi-supervised learning has proven to be a powerful paradigm for leveraging unlabeled data to mitigate the reliance on large labeled datasets. In this work, we unify the current dominant approaches for semi-supervised learning to produce a new algorithm, MixMatch, that works by guessing low-entropy labels for data-augmented unlabeled examples and mixing labeled and unlabeled data using MixUp. We show that MixMatch obtains state-of-the-art results by a large margin across many datasets and labeled data amounts. For example, on CIFAR-10 with 250 labels, we reduce error rate by a factor of 4 (from 38
Homepage: https://arxiv.org/abs/1905.02249
Related Software: CIFAR; PyTorch; Python; SGDR; ImageNet; ReMixMatch; FixMatch; Scikit; mixup; RandAugment; FlexMatch; SemiBoost; S4L; Adam; t-SNE; TensorFlow; LAMDA-SSL; AutoAugment; AugMix; TADAM
Referenced in: 7 Publications

Referencing Publications by Year