/reliable_ai

riai project for neural network verification

Primary LanguagePythonMIT LicenseMIT

RIAI 2019 Course Project

Folder structure

In the directory code you can find 2 files. File networks.py contains encoding of fully connected and convolutional neural network architectures as PyTorch classes. The architectures extend nn.Module object and consist of standard PyTorch layers (Linear, Flatten, ReLU, Conv2d). Please note that first layer of each network performs normalization of the input image. File verifier.py contains a template of verifier. Loading of the stored networks and test cases is already implemented in main function. If you decide to modify main function, please ensure that parsing of the test cases works correctly. Your task is to modify analyze function by building upon DeepZ convex relaxation. Note that provided verifier template is guaranteed to achieve 0 points (by always outputting not verified).

In folder mnist_nets you can find 10 neural networks (5 fully connected and 5 convolutional). These networks are loaded using PyTorch in verifier.py. In folder test_cases you can find 10 subfolders. Each subfolder is associated with one of the networks, using the same name. In a subfolder corresponding to a network, you can find 2 test cases for this network. As explained in the lecture, these test cases are not part of the set of test cases which we will use for the final evaluation.

Setup instructions

We recommend you to install Python virtual environment to ensure dependencies are same as the ones we will use for evaluation. To evaluate your solution, we are going to use Python 3.6.9. You can create virtual environment and install the dependencies using the following commands:

$ virtualenv venv --python=python3.6
$ source venv/bin/activate
$ pip install -r requirements.txt

Running the verifier

We will run your verifier from code directory using the command:

$ python verifier.py --net {net} --spec ../test_cases/{net}/img{test_idx}_{eps}.txt

In this command, {net} is equal to one of the following values (each representing one of the networks we want to verify): fc1, fc2, fc3, fc4, fc5, conv1, conv2, conv3, conv4, conv5. test_idx is an integer representing index of the test case, while eps is perturbation that verifier should certify in this test case.

To test your verifier, you can run for example:

$ python verifier.py --net fc1 --spec ../test_cases/fc1/img0_0.06000.txt

To evaluate the verifier on all networks and sample test cases, we provide the evaluation script. You can run this script using the following commands:

chmod +x evaluate
./evaluate