dtControl: Decision Tree Learning Algorithms for Controller Representation
This work addresses the need for smaller and more explainable controller representations in correct-by-construction synthesis, though it is incremental as it builds on existing decision tree techniques.
The authors tackled the problem of representing provably-correct controllers concisely by developing dtControl, a tool that uses decision tree learning algorithms, resulting in extremely efficient decision trees with a single-digit number of nodes on 5 out of 10 case studies.
Decision tree learning is a popular classification technique most commonly used in machine learning applications. Recent work has shown that decision trees can be used to represent provably-correct controllers concisely. Compared to representations using lookup tables or binary decision diagrams, decision trees are smaller and more explainable. We present dtControl, an easily extensible tool for representing memoryless controllers as decision trees. We give a comprehensive evaluation of various decision tree learning algorithms applied to 10 case studies arising out of correct-by-construction controller synthesis. These algorithms include two new techniques, one for using arbitrary linear binary classifiers in the decision tree learning, and one novel approach for determinizing controllers during the decision tree construction. In particular the latter turns out to be extremely efficient, yielding decision trees with a single-digit number of decision nodes on 5 of the case studies.