Tree Neural Networks in HOL4
This work provides a tool for integrating neural networks into formal verification systems, but it is incremental as it applies an existing method to a new domain.
The authors implemented tree neural networks in the HOL4 proof assistant to approximate functions on formulas, achieving performance comparable to other machine learning predictors on tasks like evaluating arithmetical expressions and estimating propositional formula truth.
We present an implementation of tree neural networks within the proof assistant HOL4. Their architecture makes them naturally suited for approximating functions whose domain is a set of formulas. We measure the performance of our implementation and compare it with other machine learning predictors on the tasks of evaluating arithmetical expressions and estimating the truth of propositional formulas.