Weighted Graph-Based Signal Temporal Logic Inference Using Neural Networks
Nasim Baharisangari, Kazuma Hirota, Ruixuan Yan, A. Agung Julius, Zhe Xu · IEEE Control Systems Letters · 2021
Extracting spatial-temporal knowledge from data is useful in many applications. It is important that the obtained knowledge is human-interpretable and amenable to formal analysis. In this letter, we propose a method that trains neural networks to learn spatial-temporal properties in the form ofweighted graph-based signal temporal logic(w-GSTL) formulas. For learning w-GSTL formulas, we introduce aflexiblew-GSTLformula structurein which the user’s preference can be applied in the inferred w-GSTL formulas. In the proposed framework, each neuron of the neural networks corresponds to a subformula in aflexiblew-GSTLformula structure. We initially train a neural network to learn the w-GSTL operators, and then train a second neural network to learn the parameters in aflexiblew-GSTLformula structure. We use a COVID-19 dataset and a rain prediction dataset to evaluate the performance of the proposed framework and algorithms. We compare the performance of the proposed framework with three baseline classification methods including K-nearest neighbors, decision trees, support vector machine, and artificial neural networks. The classification accuracy obtained by the proposed framework is comparable with the baseline classification methods.