Picture for Peter König

Peter König

Balancing stability and plasticity in continual learning: the readout-decomposition of activation change (RDAC) framework

Add code
Oct 10, 2023
Viaarxiv icon

Diagnosing Catastrophe: Large parts of accuracy loss in continual learning can be accounted for by readout misalignment

Add code
Oct 09, 2023
Viaarxiv icon

Fast Concept Mapping: The Emergence of Human Abilities in Artificial Neural Networks when Learning Embodied and Self-Supervised

Add code
Feb 03, 2021
Figure 1 for Fast Concept Mapping: The Emergence of Human Abilities in Artificial Neural Networks when Learning Embodied and Self-Supervised
Figure 2 for Fast Concept Mapping: The Emergence of Human Abilities in Artificial Neural Networks when Learning Embodied and Self-Supervised
Figure 3 for Fast Concept Mapping: The Emergence of Human Abilities in Artificial Neural Networks when Learning Embodied and Self-Supervised
Figure 4 for Fast Concept Mapping: The Emergence of Human Abilities in Artificial Neural Networks when Learning Embodied and Self-Supervised
Viaarxiv icon

Enhancing Traffic Scene Predictions with Generative Adversarial Networks

Add code
Sep 24, 2019
Figure 1 for Enhancing Traffic Scene Predictions with Generative Adversarial Networks
Figure 2 for Enhancing Traffic Scene Predictions with Generative Adversarial Networks
Figure 3 for Enhancing Traffic Scene Predictions with Generative Adversarial Networks
Figure 4 for Enhancing Traffic Scene Predictions with Generative Adversarial Networks
Viaarxiv icon

Further advantages of data augmentation on convolutional neural networks

Add code
Jun 26, 2019
Figure 1 for Further advantages of data augmentation on convolutional neural networks
Figure 2 for Further advantages of data augmentation on convolutional neural networks
Viaarxiv icon

Learning robust visual representations using data augmentation invariance

Add code
Jun 11, 2019
Figure 1 for Learning robust visual representations using data augmentation invariance
Figure 2 for Learning robust visual representations using data augmentation invariance
Viaarxiv icon

Data augmentation instead of explicit regularization

Add code
Aug 01, 2018
Figure 1 for Data augmentation instead of explicit regularization
Figure 2 for Data augmentation instead of explicit regularization
Figure 3 for Data augmentation instead of explicit regularization
Figure 4 for Data augmentation instead of explicit regularization
Viaarxiv icon

Do deep nets really need weight decay and dropout?

Add code
Jul 12, 2018
Figure 1 for Do deep nets really need weight decay and dropout?
Figure 2 for Do deep nets really need weight decay and dropout?
Viaarxiv icon