Reproducing Label-Free XAI

Original Code Author: Jonathan Crabbé ([email protected])

This repository contains the reproduction of LFXAI, a framework to explain the latent representations of unsupervised black-box models with the help of usual feature importance and example-based methods. The reproduction is used for a reproducability study for the ML Reproducibility Challenge 2022. This repository uses the original code of Jonathan Crabbé (found at https://github.com/vanderschaarlab/Label-Free-XAI) with some adjustments and additions. For more details, please read the original ICML 2022 paper: 'Label-Free Explainability for Unsupervised Models'.

1. Installation

From repository:

1. Clone the repository
2. Create a new virtual environment with Python 3.8
3. Run the following command from the repository folder:

pip install .

When the packages are installed, you are ready to explain unsupervised models.

2. Toy example

Bellow, you can find a toy demonstration where we compute label-free feature and example importance with a MNIST autoencoder. The relevant code can be found in the folder explanations.

import torch
from pathlib import Path
from torchvision.datasets import MNIST
from torch.utils.data import DataLoader, Subset
from torchvision import transforms
from torch.nn import MSELoss
from captum.attr import IntegratedGradients

from lfxai.models.images import AutoEncoderMnist, EncoderMnist, DecoderMnist
from lfxai.models.pretext import Identity
from lfxai.explanations.features import attribute_auxiliary
from lfxai.explanations.examples import SimplEx

# Select torch device
device = torch.device("cuda") if torch.cuda.is_available() else torch.device("cpu")

# Load data
data_dir = Path.cwd() / "data/mnist"
train_dataset = MNIST(data_dir, train=True, download=True)
test_dataset = MNIST(data_dir, train=False, download=True)
train_dataset.transform = transforms.Compose([transforms.ToTensor()])
test_dataset.transform = transforms.Compose([transforms.ToTensor()])
train_loader = DataLoader(train_dataset, batch_size=100)
test_loader = DataLoader(test_dataset, batch_size=100, shuffle=False)

# Get a model
encoder = EncoderMnist(encoded_space_dim=10)
decoder = DecoderMnist(encoded_space_dim=10)
model = AutoEncoderMnist(encoder, decoder, latent_dim=10, input_pert=Identity())
model.to(device)

# Get label-free feature importance
baseline = torch.zeros((1, 1, 28, 28)).to(device) # black image as baseline
attr_method = IntegratedGradients(model)
feature_importance = attribute_auxiliary(encoder, test_loader,
                                         device, attr_method, baseline)

# Get label-free example importance
train_subset = Subset(train_dataset, indices=list(range(500))) # Limit the number of training examples
train_subloader = DataLoader(train_subset, batch_size=500)
attr_method = SimplEx(model, loss_f=MSELoss())
example_importance = attr_method.attribute_loader(device, train_subloader, test_loader)

3. Reproducing our results

MNIST experiments

In the experiments folder, run the following script

python -m mnist --name experiment_name

where experiment_name can take the following values:

experiment_name	description
consistency_features	Consistency check for label-free feature importance (paper Section 4.1)
consistency_examples	Consistency check for label-free example importance (paper Section 4.1)
roar_test	ROAR test for label-free feature importance (paper Appendix C)
pretext	Pretext task sensitivity use case (paper Section 4.2)
disvae	Challenging assumptions with disentangled VAEs (paper Section 4.3)

Pretext Use Case

Specifically for the pretext experiment, we made multiple additions, such that the following syntax can now be used:

python -m mnist --name pretext --feature_attr_method feature_method --example_attr_method example_method --pretrained

Where --pretrained uses pre-trained model checkpoints (only works if these are saved) to prevent repeated fitting on the same dataset. The other two command line arguments are detailed below, the default is highlighted in italics.

feature_method	example_method
IntegratedGradients	InfluenceFunctions
GradientShap	TracIn
Saliency	SimplEx
Random	DKNN

ECG5000 experiments

Run the following script

python -m ecg5000 --name experiment_name

where experiment_name can take the following values:

experiment_name	description
consistency_features	Consistency check for label-free feature importance (paper Section 4.1)
consistency_examples	Consistency check for label-free example importance (paper Section 4.1)

CIFAR10 experiments

Run the following script

python -m cifar10 --name experiment_name

where experiment_name can take the following values:

experiment_name	description
consistency_features	Consistency check for label-free feature importance (paper Section 4.1)
consistency_examples	Consistency check for label-free example importance (paper Section 4.1)

dSprites experiment

Run the following script

python -m dsprites

The experiment needs several hours to run since several VAEs are trained.