Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/google-deepmind/jax_verify

Neural network verification in JAX
https://github.com/google-deepmind/jax_verify

Last synced: 3 months ago
JSON representation

Neural network verification in JAX

Awesome Lists containing this project

README 

        
# jax_verify: Neural Network Verification in JAX



[![tests status](https://travis-ci.com/deepmind/jax_verify.svg?branch=master)](https://travis-ci.com/deepmind/jax_verify)

[![docs: latest](https://img.shields.io/badge/docs-stable-blue.svg)](https://jax-verify.readthedocs.io)



Jax_verify is a library containing JAX implementations of many widely-used neural network verification techniques.



## Overview



If you just want to get started with using jax_verify to verify your neural

networks, the main thing to know is we provide a simple, consistent interface

for a variety of verification algorithms:



```python

output_bounds = jax_verify.verification_technique(network_fn, input_bounds)

```



Here, `network_fn` is any JAX function, `input_bounds` define bounds over

possible inputs to `network_fn`, and `output_bounds` will be the computed bounds

over possible outputs of `network_fn`. `verification_technique` can be one of

many algorithms implemented in `jax_verify`, such as `interval_bound_propagation`

or `crown_bound_propagation`.



The overall approach is to use JAX’s powerful [program transformation system](https://jax.readthedocs.io/en/latest/notebooks/Writing_custom_interpreters_in_Jax.html), 

which allows us to analyze general network structures defined by `network_fn`

and then to define corresponding functions for calculating

verified bounds for these networks.



## Verification Techniques



The methods currently provided by `jax_verify` include:



* Functional Lagrangian [Berrada et al 2021](https://arxiv.org/abs/2102.09479)

* SDP-FO (first-order SDP verification, [Dathathri et al 2020](https://arxiv.org/abs/2010.11645))

* Non-convex ([Bunel et al 2020](https://arxiv.org/abs/2010.14322))

* Interval Bound Propagation ([Gowal et al 2018](https://arxiv.org/pdf/1810.12715.pdf), [Mirman et al 2018](http://proceedings.mlr.press/v80/mirman18b/mirman18b.pdf))

* Backward Lirpa bounds such as CAP ([Wong and Kolter 2017](https://arxiv.org/pdf/1711.00851.pdf)), FastLin([Weng et al 2018](https://arxiv.org/pdf/1804.09699.pdf)) or CROWN ([Zhang et al 2018](https://arxiv.org/pdf/1811.00866.pdf))

* Forward Lirpa bounds ([Xu et al 2020](https://arxiv.org/pdf/2002.12920.pdf))

* CROWN-IBP ([Zhang et al 2019](https://arxiv.org/abs/1906.06316))

* Planet (also known as the "LP" or "triangle" relaxation, [Ehlers 2017](https://arxiv.org/abs/1705.01320)), currently using [CVXPY](https://github.com/cvxgrp/cvxpy) as the LP solver

* MIP encoding ([Cheng et al 2017](https://arxiv.org/pdf/1705.01040.pdf), [Tjeng et al 2019](https://arxiv.org/pdf/1711.07356.pdf))



## Installation



**Stable**: Just run `pip install jax_verify` and you can `import jax_verify` from any of your Python code.



**Latest**: Clone this directory and run `pip install .` from the directory root.



## Getting Started



We suggest starting by looking at the minimal examples in the `examples/` directory.

For example, all the bound propagation techniques can be run with the `run_boundprop.py` script:



```bash

cd examples/

python3 run_boundprop.py --boundprop_method=interval_bound_propagation

```



For documentation, please refer to the [API reference page](https://jax-verify.readthedocs.io/en/latest/api.html).



## Notes



Contributions of additional verification techniques are very welcome. Please open

an issue first to let us know.



## License



All code is made available under the Apache 2.0 License.

Model parameters are made available under the Creative Commons Attribution 4.0

International (CC BY 4.0) License.

See https://creativecommons.org/licenses/by/4.0/legalcode for more details.



## Disclaimer



This is not an official Google product.