Dear participants,
The first month of the competition is over. During this month 1,000 Participants enrolled in the challenge and submitted more than 1,350 submissions!
The public test set is relatively small (106 Segments), the first places are already reaching high scores, but be careful not to overfit the small public test set and assure that your model can generalize well. We encourage you to keep on and climb up (carefully) at the leaderboard.
In this kind of competition, sometimes the transition from the public phase to the private phase changes the leaderboard drastically. In the first MAFAT challenge, the winning 2nd place wasn’t even in the top 10 at the end of the public phase.
Also, remember that in this competition you are allowed to retrain your model during the private phase and you will have at the end of the public phase the labels of the public test set and also the full tracks of the public test set segments to use in your retrain.
For those who just started, one of the participants published an interesting blog-post in Medium about the competition and the dataset.
Good luck,
MAFAT Challenge Team
Hello,
I am a second-year undergraduate student working on a university-mandated academic project focused on radar signal processing and learning-based analysis techniques.
Our project requires real-world radar data, and the MAFAT Radar Challenge dataset is particularly well-suited due to its structure and relevance. We intend to use the dataset strictly for academic and research purposes and will fully comply with all licensing and usage conditions set by MAFAT.
I would like to ask whether students can request access to the dataset for academic projects, and if so, what the correct procedure is (registration, approval, or any required documentation). If access is restricted, guidance on alternative datasets or official channels would also be appreciated.
Thank you for maintaining and supporting such valuable resources for the research and student community.
Hanna
Posted by: Hanna_Susan_Eldo @ Feb. 3, 2026, 9:14 a.m.