Weighted Finite Automata Learning Using Tensor Train

The repository contains experiments for approximating function with WFA using Tensor Train decomposition and spectral learning algorithm.

File experiments.ipynb runs and compares 2 algorithms:

• spectral learning algorithm
• Tensor Train decomposition and further automata extraction

Requirements

pip install -r requirements.txt

Run experiments

Use experiments.ipynb to produce experiments results. Works in Google Colab.

Results summary