Learning Broadcast Protocols with LeoParDS

Summary

LeoParDS is a new tool for learning broadcast protocols (BPs) from a set of positive and negative example traces. It is the first tool that enables learning of a distributed computational model in a parameterized setting, i.e., with a parametric number of processes running the BP concurrently. We describe the tool along a running example, discuss some implementation details, and present experimental results on randomly generated BPs.