Process Mining
Reliable process mining
While there are mature process mining techniques, deriving trustworthy knowledge through their application requires skill, especially considering that commercial tools often only implement directly-follows graphs (the lack of adoption of inductive or split mining in commercial tools could point to a general limitation in itself). Additionally, in real world projects, data processing requires much effort and influences the obtainable knowledge, but existing techniques generally abstract from log creation. We develop tools that provide end to end support for process mining and that lower the application barrier for inexperienced users.
Intelligent process mining
Existing approaches for predictive process monitoring predominantly use ML and have to be trained from scratch for a specific context, i.e., for each dataset. Reusable ML models that encode general process knowledge could serve as a basis for predictive process monitoring, limiting the training efforts. We investigate the integration of these techniques in other BPM areas, such as process modelling, process discovery, conformance checking, etc.
CSIRO Mission Alignment
Our research aligns with the vision of the Trusted AgriFood Exports Mission.