ecmlpkdd.org

ECML-PKDD LOGO

Discovery Challenge

We welcome proposals for the Discovery Challenge of ECML PKDD 2025. Each year ECML PKDD hosts several challenges in order to promote research and evaluate machine learning approaches in real-world applications. The Discovery Challenge is open to academic and industrial institutions as well as to non-profit organizations. Each selected challenge will have a dedicated session in the conference to present the solutions, and will be invited to submit one paper to the ECML/PKDD workshop proceedings that describes the challenge and the winning solution(s). Moreover, each competition should select and review the two top solutions papers to be submitted to the ECML/PKDD workshop proceedings.

Key Dates & Deadlines

Submission

Submission Deadline

2025-03-24

Author Notification

2025-03-31

*All deadlines expire on 23:59 AoE

Competition Time Frame

Start

2025-04-21

End

2025-06-30

Publish Results

2025-07-08

Submit Papers for Proceedings

2025-07-31

*All deadlines expire on 23:59 AoE

VIEW OPTIONS

Proposal Submission

Paper format

Contact

Proposal Submission

The proposals should cover the end to end organization of a challenge. The proposal should be 2 to max 4 pages and should contain the following information:

  • Problem description: Why is the problem challenging and interesting? What is the real-world application and impact (research/societal, etc.)? Potential research impact?
  • Details on the exact tasks and the data (size, availability of meta data)
  • Privacy and ethical concerns and considerations related to the problem, tasks and data.
  • Evaluation: Describe very precisely how you will perform the evaluation, and any other policies. For example, what type of metrics you will use and how you will ensure a fair and rigid evaluation, are there just quantitative tasks or also qualitative evaluations and if so by whom, and how does this lead to the final ranking (pass/fail criteria, scores etc).
  • Communication plan: How do you plan to communicate, market and promote the challenge? How do you drive the volume submissions? Are you connected to particular communities? How do you ensure the challenge is made accessible for a a wide variety of participants, so that everyone could attend? How do you plan to communicate the results?
  • Technical details: The platform you will use to host the challenge, discussion areas, leaderboard, for example codabench.org or others. Baselines and code for data accessing that you will provide to participants. Will these be shared through github?
  • Reporting requirements for the participants’ submissions, like for example, a report, code etc. and are these must have requirements or optional, pass/fail or graded etc. ECML PKDD can offer facilities to upload papers (CMT).
  • Timeline for the challenge. Note that you should respect the overall time frame mentioned above.
  • Awards: The type of tangible or intangible awards/prizes you will consider for the winning solutions.
  • Details of the organizers of the challenge (Short CVs, Contact Details).

For reference, see the 2024 competitions at https://ecmlpkdd.org/2024/program-discovery-challenge/

Submission site: https://cmt3.research.microsoft.com/ECMLPKDD2025

Please submit your proposal as: Create new submission and select the track: Discovery Challenge.

Paper format

Discovery Challenge papers (describing the top systems) must be written in English, have a maximum of 8 pages. We suggest the papers be prepared and submitted in the LNCS format (https://www.springer.com/gp/computer-science/lncs/conference-proceedings-guidelines). Springer LNCS Template Download. In case of questions, do not hesitate to contact the Discovery Challenge chairs (mail list below).

Contact

Please send your submission through CMT. For further questions and information please contact the Discovery Challenge chairs, Carlos Ferreira (Polytechnic Institute of Porto & INESC TEC), Peter van der Putten (Leiden University & Pegasystems) and Rui Camacho (University of Porto & INESC TEC) through the following mailing list: ecml-pkdd-2025-discovery-challenge-chairs@googlegroups.com