Transfer learning for jet energy scales
Objectives:
- Include (in-situ) data information into the calibration of jet energy scales (JES) at an early ML-based stage, instead of the average in-situ correction currently applied at the last step of the calibration chain
- Improve the jet energy resolution due to considering the correlations among the variables relevant for the JES calibration and to integrating the in-situ constraints at the step of the training of the ML algorithm
- Combination of supervised and unsupervised methods for solving inverse problems
Expected Results:
- Reduction of simulation bias in the jet energy calibration and in solving inverse problems
- Combination of data and theoretical knowledge
- Reduced jet energy calibration uncertainties
- Improved jet energy resolution
