Skip to main content

SMT

MLDiff
MLDiff provides a novel approach to comparing learned classifiers by translating them into SMT formulas and systematically analyzing their decision boundaries.
VS Code Extension for SMT-LIB (Z3)
A Visual Studio Code extension for the SMT Z3 theorem prover. The extension provides a user-friendly interface to run the Z3 solver in the VS Code editor without installing Z3 separately.
SMT-LIB LSP
A language server for the SMT-LIB language, providing features like syntax highlighting, autocompletion, and error checking for SMT-LIB files in various editors. It is based on Langium workbench and integrates with the Z3 SMT solver.
On the Comparison of Learned Classifiers
On Writing SMT-LIB Scripts: Metrics and a New Dataset
Towards Comparing Learned Classifiers