Skip to content

The rep for the RotateNetworks in ICPR18, Beijing, China.

License

Notifications You must be signed in to change notification settings

xialeiliu/RotateNetworks

Repository files navigation

Rotate your Networks: Better Weight Consolidation and Less Catastrophic Forgetting

The paper has been accepted as oral presentation in ICPR 2018. An arXiv pre-print version is available. Project page is available.

Abstract

In this paper we propose an approach to avoiding catastrophic forgetting in sequential task learning scenarios. Our technique is based on a network reparameterization that approximately diagonalizes the Fisher Information Matrix of the network parameters. This reparameterization takes the form of a factorized rotation of parameter space which, when used in conjunction with Elastic Weight Consolidation (which assumes a diagonal Fisher Information Matrix), leads to significantly better performance on lifelong learning of sequential tasks. Experimental results on the MNIST, CIFAR-100, CUB-200 and Stanford-40 datasets demonstrate that we significantly improve the results of standard elastic weight consolidation, and that we obtain competitive results when compared to the state-of-the-art in lifelong learning without forgetting.

Requirements

The latest version of Tensorflow and Jupyter Notebook.

Demo on MNIST dataset

Run Notebook with different settings.

Download the pre-trained ImageNet VGG-16 modeland put it in this foleder.

Download the CUB-200 Birds dataset.

Change the data path in this file.

Run script for both EWC and our appraoch R-EWC.

./Birds_vgg16/REWC_Birds_vgg16.sh
./Birds_vgg16/EWC_Birds_vgg16.sh

Evaluation

We report the average accuracy as the evaluation metric. Note that all our experiments are based on single-head implementation. You could also test on multi-head setting by changing the test parameters.

Useful functions

You could find useful functions on different datasets and networks here.

An example on fully connected layer

model

Citation

Please cite our paper if you are inspired by the idea.

@inproceedings{xialei2018forgetting,
title={Rotate your Networks: Better Weight Consolidation and Less Catastrophic Forgetting},
author={Liu, Xialei and Masana, Marc and Herranz, Luis and Van de Weijer, Joost and Lopez, Antonio M and Bagdanov, Andrew D},
booktitle={International Conference on Pattern Recognition (ICPR)},
url = {https://github.com/xialeiliu/RotateNetworks},
year={2018}
}

About

The rep for the RotateNetworks in ICPR18, Beijing, China.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published