We present a new and simple approach to active automata learning . Instead of focusing on equivalence of observations, $L^{\#}$ takes a different perspective . It triesto establish apartness, a constructive form of inequality . Experiments with a prototype implementation, written in Rust, suggest that $L${\#$ outperforms existing algorithms . We hope to find that adaptive distinguishing sequences can be naturally integrated to boost the performance of $L#$ in practice .

Author(s) : Frits Vaandrager, Bharat Garhewal, Jurriaan Rot, Thorsten WiƟmann

Links : PDF - Abstract

Code :

Keywords : approach - active - automata - learning - apartness -

Leave a Reply

Your email address will not be published. Required fields are marked *