no code implementations • 29 Sep 2023 • Lukas Meiner, Jens Mehnert, Alexandru Paul Condurache
In particular, we are able to instantly drop 46. 72% of FLOPs while only losing 1. 25% accuracy by just swapping the convolution modules in a ResNet34 on CIFAR-10 for our HASTE module.
no code implementations • 2 Mar 2023 • Matthias Rath, Alexandru Paul Condurache
We then address the problem of incorporating multiple desired invariances into a single network.
no code implementations • 17 May 2022 • Paul Wimmer, Jens Mehnert, Alexandru Paul Condurache
By freezing weights, the number of trainable parameters is shrunken which reduces gradient computations and the dimensionality of the model's optimization space.
no code implementations • CVPR 2022 • Paul Wimmer, Jens Mehnert, Alexandru Paul Condurache
Finally, we show that advances of IP are due to improved trainability and superior generalization ability.
no code implementations • 8 Feb 2022 • Matthias Rath, Alexandru Paul Condurache
We demonstrate the improved sample complexity on the Rotated-MNIST, SVHN and CIFAR-10 datasets where rotation-invariant-integration-based Wide-ResNet architectures using monomials and weighted sums outperform the respective baselines in the limited sample regime.
no code implementations • 21 Aug 2020 • Julia Lust, Alexandru Paul Condurache
This survey connects the three fields within the larger framework of investigating the generalization performance of machine learning methods and in particular DNNs.
no code implementations • 30 Jun 2020 • Matthias Rath, Alexandru Paul Condurache
One promising approach, inspired by the success of convolutional neural networks in computer vision tasks, is to incorporate knowledge about symmetric geometrical transformations of the problem to solve that affect the output in a predictable way.
no code implementations • 20 Apr 2020 • Julia Lust, Alexandru Paul Condurache
Deep neural networks (DNNs) are vulnerable to adversarial examples and other data perturbations.
no code implementations • 20 Apr 2020 • Matthias Rath, Alexandru Paul Condurache
In this contribution, we show how to incorporate prior knowledge to a deep neural network architecture in a principled manner.
no code implementations • ECCV 2018 • Benjamin Coors, Alexandru Paul Condurache, Andreas Geiger
Omnidirectional cameras offer great benefits over classical cameras wherever a wide field of view is essential, such as in virtual reality applications or in autonomous robots.