8.8. Advanced Topics

• 8.8.1. Rest Of Event
  • Overview
  • Basic usage
  • ROE particle type hypothesis
  • Charged PID priors
  • Selection cut based method
  • Accessing ROE particles
  • ROE masks
  • Nested ROE
  • Load ROE as a particle
  • Create ROE using MCParticles
  • MVA based cleaning

• 8.8.2. How to Veto
  • Introduction
  • Overview of the workflow
  • Examples

• 8.8.3. Flavor Tagger
  • Flavor Tagging Principle
  • Flavor Tagger Algorithm
  • Using the FlavorTagger
  • Saving to nTuples
  • Efficiency Calculation and Validation Plots
  • Tutorials
  • Functions
    • flavorTagger()

• 8.8.4. Transformer Flavor Tagger (TFlaT)
  • Principle
  • Algorithm
  • How to use
  • Tutorials
  • Functions
    • flavorTagger()

• 8.8.5. Deep Flavor Tagger
  • Principle
  • Algorithm
  • How to use
  • Functions
    • DeepFlavorTagger()

• 8.8.6. Charm Flavor Tagger
  • charmFlavorTagger()
  • Tutorials

• 8.8.7. Full event interpretation
  • Algorithm overview
    • Physics
    • Key Performance Indicators
    • Hadronic, Semileptonic and Inclusive Tagging
    • Hierarchical Approach
  • Applying the FEI
  • FEI and the conditions database
  • Sphinx documentation
    • DecayChannel
      • DecayChannel.daughters
      • DecayChannel.decayModeID
      • DecayChannel.decayString
      • DecayChannel.extraPathSpec
      • DecayChannel.label
      • DecayChannel.mvaConfig
      • DecayChannel.name
      • DecayChannel.pi0veto
      • DecayChannel.preCutConfig
    • FeiConfiguration
      • FeiConfiguration.cache
      • FeiConfiguration.externTeacher
      • FeiConfiguration.legacy
      • FeiConfiguration.monitor
      • FeiConfiguration.monitoring_path
      • FeiConfiguration.prefix
      • FeiConfiguration.roundMode
      • FeiConfiguration.training
    • FeiState
      • FeiState.excludelists
      • FeiState.fsplists
      • FeiState.path
      • FeiState.plists
      • FeiState.stage
    • MVAConfiguration
      • MVAConfiguration.config
      • MVAConfiguration.method
      • MVAConfiguration.sPlotVariable
      • MVAConfiguration.spectators
      • MVAConfiguration.target
      • MVAConfiguration.variables
    • Particle
      • Particle.addChannel()
      • Particle.channels
      • Particle.daughters
      • Particle.extraPathSpec
      • Particle.identifier
      • Particle.label
      • Particle.mvaConfig
      • Particle.name
      • Particle.postCutConfig
      • Particle.preCutConfig
    • PostCutConfiguration
      • PostCutConfiguration.bestCandidateCut
      • PostCutConfiguration.value
    • PreCutConfiguration
      • PreCutConfiguration.bestCandidateCut
      • PreCutConfiguration.bestCandidateMode
      • PreCutConfiguration.bestCandidateVariable
      • PreCutConfiguration.bkgSamplingFactor
      • PreCutConfiguration.noBackgroundSampling
      • PreCutConfiguration.noSignalSampling
      • PreCutConfiguration.userCut
      • PreCutConfiguration.vertexCut
    • do_trainings()
    • get_ccbarLambdaC_channels()
    • get_default_channels()
    • get_mode_names()
    • get_path()
    • get_stages_from_particles()
    • get_unittest_channels()
    • save_summary()
  • Code structure
    • config.py
    • default_channels.py
    • core.py
  • Training the FEI
    • Basic Workflow (training)
    • distributed.py
    • Restarting a training
  • FEI Training on the Grid
    • Software and Environment Setup
    • General Workflow Concept
    • Technical details
      • settings.json
      • B_generic_train.py
      • FEIAnalysisSummaryTask and FEIAnalysisTask
        • The Problem of Too Long Runtimes
      • MergeOutputsTask
      • FEITrainingTask
      • PrepareInputsTask
      • Further Comments on fei_grid_workflow.py
    • Tips and Tricks
    • Possible Improvements
      • Potential Improvements Following gbasf2 Development
  • Troubleshooting
    • Crash in the Neurobayes Library
    • Running FEI on converted Belle MC outside of KEK
    • Speeding up the FEI
  • Resources, Publications etc.

• 8.8.8. ModeSelector
  • When to use ModeSelector
  • Algorithm
    • Two-stage network
    • D* veto
  • Outputs
  • Best candidate selection
  • How to use
  • Training
  • Functions
    • modeSelector()
    • addDstarVeto()

• 8.8.9. Graph-based Full Event Interpretation
  • Model training
  • Applying the model to data
  • Code documentation
    • Core modules
      • graFEI()
      • lcaSaver()
    • Other modules and functions
      • load_config()
      • GraFEIIgniteTrainer
      • create_dataloader_mode_tags()
      • populate_avail_samples()
      • preload_root_data()
      • compute_cosTheta()
      • compute_doca()
      • compute_edge_features()
      • GraphDataSet
      • EdgeLayer
      • GlobalLayer
      • NodeLayer
      • GraFEIModel
      • InvalidLCAMatrix
      • Node
      • lca_to_adjacency()
      • select_good_decay()
      • PerfectEvent
      • PerfectLCA
      • PerfectMasses
      • MultiTrainLoss
      • normalize_features()
      • is_valid_tree()
      • masses_to_classes()
      • LCASaverModule
      • get_object_list()
      • pdg_to_lca_converter()
      • write_hist()
      • GraFEIModule

• 8.8.10. EventShape
  • Introduction
  • What is the difference between event shape and continuum suppression?
  • How to use eventShape
  • Event shape variables

• 8.8.11. Continuum suppression
  • Example usage
  • Continuum Suppression variables
  • Deep Continuum Suppression
    • Adversarial Networks

• 8.8.12. Basf2 MVA interface

• 8.8.13. PID Prior Probabilities
  • Training
    • Named Arguments
    • Required Arguments
  • Evaluation
    • Import
    • Initialize
    • Prior Calculation
    • Getting the Priors
    • Getting the Posteriors

• 8.8.14. ksSelector
  • V0Selector
  • LambdaVeto
  • Tutorials
  • ksSelector Functions
    • LambdaVeto_Training()
    • LightGBM2ONNX()
    • V0Selector_Training()
    • add_default_ks_Selector_aliases()
    • add_variable_collection()
    • ksSelector()

• 8.8.15. PID Calibration Weights
  • Preparing Data Samples for Training
  • Training the Weights
    • Positional Arguments
    • Named Arguments
  • Applying Weights to Data for Performance Analysis
  • PID calibration weights on the basf2 path
  • pidDataUtils Functions
    • apply_weights()
    • compute_abl_lrs()
    • compute_contrib()
    • compute_det_lrs()
    • cut_particles()
    • make_binary_lrs()
    • make_bins()
    • make_columns()
    • make_contrib()
    • make_h5()
    • make_labels()
    • make_lrs()
    • make_pid()
    • make_pid_abl()
    • make_pid_det()
    • merge_h5s()
    • produce_analysis_df()
    • read_h5()
    • read_npz()
    • read_root()
    • root_column()
    • softmax()
    • split_h5()

• 8.8.16. Signal Embedding
  • Documentation of the scripts:
    • Named Arguments
    • Named Arguments

• 8.8.17. Offline reweighting
  • Introduction
    • Charged Particle Identification corrections
    • Full Event Interpretation corrections
    • Modules and functions
      • add_weights_to_dataframe()