Reinterpretation with Open Data: Glossary

Reinterpretation is the comparison of experimental results with the expectations of a given a theoretical model which was not already interpreted by the original analysis publication.

Reinterpretation requires running, and usually even rewriting the analysis. One can use the original analysis as it is (exact reinterpretation) or a modified version of it to optimize signal sensitivity (optimized reinterpretation.)

Open data is a great source of events, in particular for optimized reinterpretation studies.

ADL is a declarative domain specific language (DSL) that describes the physics content of a HEP analysis in a standard and unambiguous way.

ADL’s purpose is to decouple the physics logic of analyses from technical operations, and make the physics logic more accessible.

CutLang is a runtime interpreter that reads and understands the ADL syntax and runs it on events.

For up-to-date details for installing CutLang, the official documentation is the best bet.

Make sure you were able to setup CutLang via Docker and run its hello-world example.

Running CLA is the only thing that a user must know in order to work with CutLang.

The example ADL files in the /CutLang/runs/tutorials/ directory is the best way to immediately familiarize with the ADL syntax.

The ADL/CutLang allows practical and transparent implementation and optimization of an analysis for reinterpretation purposes using open data.

Optimized reinterpretation involves finding variables that discriminate signal from background; updating the event selection with cuts based on these discriminating variables; checking cutflows, variable distributions at various stages of the selection for high signal to background ratios; and finding a good final variable with the best signal to background ratio to express the analysis result.