hahtorch 0.0.5

Hebbian/Anti-Hebbian Learning for Pytorch

Figure 1: HaH block for image classification DNNs.

Hebbian/Anti-Hebbian Learning for Pytorch

Official repository for the paper entitled "Towards Robust, Interpretable Neural Networks via Hebbian/anti-Hebbian Learning: A Software Framework for Training with Feature-Based Costs". If you have questions you can contact metehancekic [at] ucsb [dot] edu.

Maintainers: WCSL Lab, Metehan Cekic, Can Bakiskan,

Dependencies

numpy==1.20.2
torch==1.10.2

How to install

The most recent stable version can be installed via python package installer "pip", or you can clone it from the git page.

pip install hahtorch

or

git clone [e-mail]:metehancekic/HaH.git

Experiments

We used CIFAR-10 image classification to show the effectiveness of our module. We train a VGG16 in a standard fashion and train another VGG16 that contains HaHblocks with layer-wise HaHCost as a supplement. Details of our experiments can be found in our recent paper

CIFAR10 Image Classification with VGG16 model as Backbone

Figure 2: HaH VGG16, our proposed architecture for HaH training, see paper for more detail.

Table 1: CIFAR10 classification: Performance of the HaH trained network against different input corruptions on the test set. For all of the adversarial attacks, we use AutoAttack which is an ensemble of parameter-free attacks, see paper for more detail.

Current Version

0.0.5

Sources

• PyPi page for the code

• Git repo for the code