Active Automata Learning Algorithms
- Angluin’s L* (extended, configurable implementation for DFA and Mealy machines)
- Direct Hypothesis Construction (Mealy)
- Maler/Pnueli (DFA, Mealy)
- Kearns/Vazirani (DFA, Mealy)
- Rivest/Schapire (DFA, Mealy)
- TTT (DFA, Mealy)
- NL* (NFA)
Be sure to also check out the blazing performance of LearnLib’s learning algorithm implementations.
Equivalence Test Approximation Algorithms
- Complete, depth-bounded exploration
- Random words
- Random walk
- Query cache
- Reuse filter
- Generic, extensible design
- Logging subsystem
The following table lists where the open-source LearnLib features differ from those of the former, closed-source version. Note that the publicly available version of the old LearnLib only contains a subset of all implemented features.
|Feature||Open-source LearnLib||Old LearnLib (public release)||Old LearnLib (internal)|
|Register Automata learning||✗||✗||✓|
|Graphical modeling tool (LearnLib studio)||✗||✓||✓|
Contact us if you are interested in a feature that currently is available in the internal version of the old LearnLib only. Maybe you are even interested in porting this feature to the new, open-source LearnLib. Your contribution would greatly be appreciated.