learningsma

A Java library and framework for learning Symbolic Mealy Automata built on top of AutomataLib and the Symbolic Automata toolkit.

Prerequisites

• Java 8 or higher
• Maven 3.6 or higher

Installing Symbolic Automata dependency

This project depends on the Symbolic Automata toolkit for symbolic finite automata (SFA) support. Clone and install it into your local Maven repository:

git clone --recursive https://github.com/lorisdanto/symbolicautomata.git /tmp/symbolicautomata
# If you omitted --recursive above, initialize submodules:
git -C /tmp/symbolicautomata submodule update --init --recursive
cd /tmp/symbolicautomata
# Use -DskipTests to speed up and, if necessary, -Dmaven.repo.local to point to a writable repo
mvn clean install -DskipTests

Building learningsma

Clone this repository and run Maven. After installing the Symbolic Automata toolkit (models module), all dependencies will be pulled from your local repository and Maven Central.

git clone https://github.com/SoftwareFoundationGroupAtKyotoU/learningsma.git
cd learningsma
mvn clean install

Dependencies

Aside from the Symbolic Automata toolkit (models module installed above), all other dependencies are pulled automatically by Maven. Notable libraries include:

• AutomataLib (via net.automatalib.distribution:automata-distribution)
• SAT4J (org.ow2.sat4j:org.ow2.sat4j.core)
• Guava (com.google.guava:guava)
• Apache Commons Lang & Collections
• JavaFX (controls & FXML)

Running examples

Several example classes are provided in src/main/java/example. For instance:

mvn exec:java -Dexec.mainClass="example.Example"

Reference

• Kengo Irie, Masaki Waga, Kohei Suenaga, "Active Learning of Symbolic Mealy Automata." ICTAC 2025.