TADAM swMATH ID: 41279 Software Authors: Boris N. Oreshkin, Pau Rodriguez, Alexandre Lacoste Description: TADAM: Task dependent adaptive metric for improved few-shot learning. Few-shot learning has become essential for producing models that generalize from few examples. In this work, we identify that metric scaling and metric task conditioning are important to improve the performance of few-shot algorithms. Our analysis reveals that simple metric scaling completely changes the nature of few-shot algorithm parameter updates. Metric scaling provides improvements up to 14 Homepage: https://arxiv.org/abs/1805.10123 Source Code: https://github.com/ElementAI/TADAM Related Software: MixMatch; AutoAugment; ImageNet; AugMix; Adam; SGDR; SpecAugment; S4L; mixup; RandAugment; ReMixMatch; CIFAR; Python; FixMatch Cited in: 0 Publications