NeurASP: Neural Networks + Answer Set Programs

NeurASP is a simple extension of answer set programs by embracing neural networks. By treating the neural network output as the probability distribution over atomic facts in answer set programs, NeurASP provides a simple and effective way to integrate sub-symbolic and symbolic computation. This repository includes examples to show

1. how NeurASP can make use of pretrained neural networks in symbolic computation and how it can improve the perception accuracy of a neural network by applying symbolic reasoning in answer set programming; and
2. how NeurASP is used to train a neural network better by training with rules so that a neural network not only learns from implicit correlations from the data but also from the explicit complex semantic constraints expressed by ASP rules.

1. Install Python 3.7 version of Anaconda according to its installation page.

2. Install clingo using the following command line. (clingo 5.3 and 5.4 are tested)

conda install -c potassco clingo

3. Install PyTorch according to the its home page. (PyTorch version 1.0.1, 1.3.0, and 1.4.0 are tested)

Clone this repo:

git clone https://github.com/zhunyoung/NeurASP
cd NeurASP

We provide 3 inference and 5 learning examples as shown below. Each example is stored in a separate folder with a readme file.

• Sudoku
• Offset Sudoku
• Toy-car

• MNIST Addition
• Shorstest Path
• Sudoku Solving
• Top-k
• Most Reliable Path