Machine-learned particle flow

MLPF reconstructs the particles produced in a collision from detector measurements. It uses a neural network to combine information from tracking detectors and calorimeters and produces a list of particle candidates for the full event. The project supports studies with the CMS detector and with the CLD and CLIC future-detector concepts.

This documentation describes the current development version of the repository. A result reported in a paper belongs to the code, data, detector setup, and model used in that study; it is not automatically a performance claim for the current branch.

Where should I start?¶

The workflow at a glance¶

detector simulation
        |
        v
tracks, calorimeter clusters, or detector hits
        |
        v
detector-specific postprocessing and data checks
        |
        v
versioned TensorFlow Dataset (TFDS)
        |
        v
MLPF training and checkpoint
        |
        v
particle predictions
        |
        v
particle, jet, missing-momentum, and deployment validation

The full workflow is available for CMS, CLD, and CLIC. New users normally do not need to generate detector simulation: published datasets and model checkpoints are available from the MLPF Hugging Face repositories.

Support levels¶

The documentation uses four labels:

• Supported: there is a maintained configuration and a tested path in this repository.

• Research workflow: the implementation is usable, but it needs more detector, model, or computing knowledge.

• Partial: only part of the end-to-end workflow is implemented.

• Planned: the work is proposed or under development and is not a current capability.

See Current capabilities for the detailed matrix.

Project links¶

• Source code

• Datasets

• Models

• Issues

• Discussions

• Documentation plan