Projects

Metadata ecosystem

OpenMAPP exists to connect multiple tools together, so more powerful physics results can be derived by leveraging the best aspects of each input. A precursor to this combination is awareness of what code implementation of published data-analyses exist, in which interpretation toolkits.

We are providing this metadata machinery for the community via a new JSON format for analysis declaration, database-linking and semantic-metadata mechanisms in the HEPData and Inspire data repositories, and associated tools (such as static analysis of declared event-topologies) for studying analysis coverage and overlaps.

Analysis interoperability

The various MC simulation and event-analysis tools on the market use different workflows, interfaces, and data formats. OpenMAPP will drive standardisation and interoperability, through use of common statistical libraries, extension and use of the community HS3 data format, and providing a multi-tool steering layer. The ADL domain-specific analysis description language is developing a workflow-management tool, and mechanisms for analysis output to be used in several OpenMAPP frameworks.

Analysis combination

Accurate statistical combination requires care not to double-count physics-model features, which can make themselves known in several observables at the same time. If the same events might be used in construction of those observables, they are not statistically independent and we need to either use the single best observable or to make use of estimated correlations between measurements to make the best combination of independent observables. OpenMAPP is extending the TACO procedure for this estimation and optimisation, extending analysis tools to implement event-level bin tagging, and validating existing appraoches such as Contur’s heuristic analysis pools. In parallel, AI and static-analysis approaches to ADL specifications will provide another check of analysis combinability.

OpenMAPP developed/associated tools

Analysis:

Statistics: