3 ms·
Yes, their first paper is on MNIST. That they get high accuracy isn't earth-shattering, but since they are doing something very different from other approaches,
by yorwba 9y ago
Yes, their first paper is on MNIST. That they get high accuracy isn't earth-shattering, but since they are doing something very different from other approaches, it's still noteworthy. The real benefit is in the generalization performance:
We then tested this network on the affNIST 4 data set, in which each example is an MNIST digit with
a random small affine transformation. Our models were never trained with affine transformations
other than translation and any natural transformation seen in the standard MNIST. An under-trained
CapsNet with early stopping which achieved 99.23% accuracy on the expanded MNIST test set
achieved 79% accuracy on the affnist test set. A traditional convolutional model with a similar
number of parameters which achieved similar accuracy (99.22%) on the expanded mnist test set but
only achieved 66% on the affnist test set.
- amelius 9y agoYes that seems like a good first test for generalization. Did they publish these images somewhere?