KG: {(chop,canbedone with, knife), (knife, islocatedat, kitchen)}. As anexample, forboth (sleep, ./,

Towards understanding learning representations: To what extent do different neural networks learn the samerepresentation.

Compared with the existing optimization analysis for Meta-Learning in FL, federated conditional stochastic optimization considers the sample of tasks.

Compared with the existing optimization analysis for Meta-Learning in FL, federated conditional stochastic optimization considers the sample of tasks.

Dashed boxes denote layers that are not always present. For the MNIST energy-based modelling experiments, we use average pooling operations since other work [19, 4] has found this to produce higher quality samples than max pooling. We note that that another commonly used feature in recent GAN and EBM architectures is Spectral Normalisation (SN) [13]. Our preliminary experiments suggested that SN was not beneficial for performance. As stated in the main text, the number and (in the case of linear combinations) the spacing of the waymarks aretreated ashyperparameters.

Density-ratio estimation via classification is a cornerstone of unsupervised learning.

Wehavetried using the same number of layers for GRAND/BLEND as in their paper.However, the test clean accuracy is low. For example, grb-cora [7] has node feature size of 302 while the original Cora dataset has node feature size of 1433.

To defend against adversarial attacks, several robust GNN models have been proposed.