Embedded Files
Friday April 30th, online, public, recorded
Friday April 30th, online, public, recorded
During this 3 ½ hour "Final Jamboree" event, we presented the challenge, its organization, summarized how the data were simulated and showed our in-house analysis for planet c.
During this 3 ½ hour "Final Jamboree" event, we presented the challenge, its organization, summarized how the data were simulated and showed our in-house analysis for planet c.
We explained which metrics were applied to the participants' entries to rank them.
We explained which metrics were applied to the participants' entries to rank them.
Finally, we announced the winners/prizes and discussed the lessons learned, the the legacy tutorial as well as heard the 3 finalist teams give their feedback!
Finally, we announced the winners/prizes and discussed the lessons learned, the the legacy tutorial as well as heard the 3 finalist teams give their feedback!
About the Roman Exoplanet Imaging Data Challenge
About the Roman Exoplanet Imaging Data Challenge
The ROMAN Exoplanet Imaging Data Challenge ran for 8 months in 2019-2020. Organized by the Turnbull Science Investigation Team (SIT), this data challenge was a unique opportunity for exoplanet scientists of all backgrounds and experience levels to get acquainted with realistic Roman CGI (coronagraphic) simulated data with a new contrast regime enabling to unveil giant planets in reflected light.
Participants had to recover the astrometry of an exoplanetary system combining precursor radial velocity data with up to 6 coronagraphic imaging epochs throughout the mission. They had to perform accurate orbital fitting and determine the mass of any planet hidden in the data. It involved PSF subtraction techniques, post-processing, and other astrophysics hurdles to overcome such as contamination sources (stellar, extragalactic, and exozodiacal light).
We are now ready to announce the results to the community, present our in-house analysis, and discuss lessons learned and paths forward to best exploit scientifically the data coming out of Roman CGI. For that, we welcome you to attend our 3 ½ hour event on Friday, April 30 2021. The event was to be publicly live-streamed on STScI Live Science Events' channel (no need to have Facebook account). However, if you would like to actively participate in the discussion and/or the Q&As, please register and we will send you a link to connect.
AGENDA (GMT-4 / EDT)
AGENDA (GMT-4 / EDT)
Video of Session 1
11:00-12:00 Presentation (Chair: Margaret Turnbull)
Welcome message & motivations for this challenge - Margaret Turnbull
Data simulations (imaging/radial velocity) and architecture of the hidden exoplanetary system - Neil Zimmerman
Organization, steps, tutorials, legacy dataset & tutorial notebooks - Julien Girard
Source identification hurdles and "in-house" analysis tour for one planet - Ell Bogat
Q&A
12:00 - 12:45 Results (Chair: Margaret Turnbull)
Prizes: what and why - Julien Girard
Metrics and astrometric analysis of the results - Ell Bogat
Results and winners - Julien Girard
Q&A
30 min BREAK
Video of Session 2
1:15 - 2:00 Feedback and Lessons learned from finalist teams (Chair: Neil Zimmerman)
Introduction of the teams and lessons learned exercise - Julien Girard
Team ▼ TRIANGLE with a focus on photometry, phase curve and albedo
Team ♦︎ DIAMOND with a focus on how to apprehend CGI reflected light data from previous experiences with ground based data
Team ⬟ PENTAGON with a focus on orbital fitting (integrating RV) & post-processing
Q&A
2:00- 2:30 Discussion & wrap up (Chairs: Sergi Hildebrandt, Julien Girard and the Team)
How to prepare to real CGI data?
The increasing role of post-processing for the real, less idealistic data and/or the most recent observing scenarios with higher fidelity model uncertainty factors.
Other CGI modes (prism spectroscopy)
The star shade rendezvous
Google Sites
Report abuse