RMDC26 Sign Up
Now open

RMDC26 is a community challenge to model simulated Roman microlensing data ahead of the mission’s Galactic Bulge Time Domain Survey. The goal is to broaden participation, standardize best practices, and spur innovation in microlensing modeling and tooling.

Roman is expected to detect tens of thousands of microlensing events during its prime mission and thousands of exoplanets, probing regions of exoplanet parameter space beyond earlier surveys.

Why participate:

  • Learn or sharpen microlensing modeling skills
  • Try and compare open‑source tools on realistic data
  • Contribute new ideas and pipelines before launch

The data in this challenge is intended to be a semi-realistic representation of the microlensing data volume and type expected from the Roman Galactic Bulge Time Domain Survey.

Our official launch will be during AAS 247, where we will host a workshop covering various aspects of the Data Challenge (specific date and time TBD).

Important Dates

Milestone Date
Challenge sign-up opens early January 2026
Kick-off workshop at AAS 247 January 4, 2026
Submission deadline April 2026
Evaluation begins May 2026
Evaluation panel discussions July 2026
Evaluation panel final reports due July 2026

Challenge Overview

Ground Rules

1) Document dependencies and CPU hours.
2) All developed code should be documented and open source.
3) One final submission per team.
4) Submissions must pass format validation with the submission tool.
5) One designated team lead/contact person.
6) Communication with experienced microlensing researchers is encouraged.

Challenge Tiers

RMDC26 will offer two challenge tiers: a beginner tier for those new to microlensing and an experienced tier for veteran microlensing researchers. Anyone can participate in either tier, but the beginner tier is not intended for entire teams with extensive microlensing modeling experience. Teams may make submissions in both tiers.

Participant Goals

1) Classify all targets.
2) Fit appropriate models to the microlensing events and tabulate the parameters derived in each case, optionally including derived physical parameters.
3) Describe the approach taken to modeling; the software and hardware used; and any innovations in modeling or classification made in response to the challenge.
4) Optional: provide plotted output for each model and posterior samples.

Submission

Submissions are validated using the microlens-submit tool. It validates and packages submissions and creates draft dossiers so you can preview how your submission will look to evaluators. It is also useful for managing your many events and solutions and tracking your progress. You can use this tool through the CLI or Python API. It installs as a Python package via pip:

# PyPI
pip install microlens-submit

# Conda-Forge
conda install microlens-submit

The microlens-submit documentation includes detailed specifications for how to create a submission. Strict adherence to the submission criteria is required, as much of the evaluation process is automated.

Evaluation

The evaluation panel will assess each team’s entry on the following criteria:

# Criterion Points
1 Accuracy of model parameters 5
2 Number of events modeled 5
3 Software/computational efficiency 5
4 Innovation 5
5 Broadening the field by bringing in new researchers 5

You can find a copy of the evaluation rubric here.

Microlensing Resources and References

Contact