Skip to content

OpportunityLab

Scholarships, grants, fellowships and internships for young Africans

Random News

OpportunityLab

Scholarships, grants, fellowships and internships for young Africans

Random News

Subscribe to our newsletter

Get the top stories delivered to your inbox every morning.

About Us

Useful Information

Vim in meis verterem menandri, ea iuvaret delectus verterem qui, nec ad ferri corpora.

Euismod nisi porta lorem mollis. Interdum velit euismod in pellentesque.

Terms & Services | Privacy Policy

Headlines

Underwater Vessel Noise Challenge 2026 on Zindi: US$5,000 for Machine Learning on South African Ocean Sound Data, Plus a ChangeNOW Summit Invitation (Open to All, Submit by 13 December)

Ephraim Akpan03 min read
Underwater Vessel Noise Challenge flyer: US$5,000 machine learning competition on Zindi, submit by 13 December 2026 Closes 13 Dec

The French Embassy in South Africa is running the Underwater Vessel Noise Challenge on Zindi, with partners including ITU’s AI for Good and the South African National Biodiversity Institute (SANBI). Participants use machine learning to detect and measure ship and boat noise in real hydrophone (underwater microphone) data from the South African coast. The total prize pool is US$5,000.

The challenge is open to all, so data scientists, students and researchers in Nigeria and across Africa can join alone or in teams of up to four. Besides leaderboard prizes, there are two US$1,000 awards for the best research write-ups, open to any participant whatever their leaderboard position.

Deadline: 13 December 2026, 11:30 UTC for both predictions and write-ups. That is 12:30 WAT, 13:30 CAT/SAST and 14:30 EAT. Results are revealed on 29 January 2027.

Key Details

  • Organiser: French Embassy in South Africa, on Zindi
  • Data: one-minute sound-energy summaries from hydrophones run by SAEON, SANBI and the SA Underwater Radiated Noise Consortium, plus AIS vessel-tracking data
  • Task: for each test window, predict the share of energy from an added vessel-sound component and an evidence score (both between 0 and 1)
  • Prize pool: US$5,000
  • Team size: up to 4
  • Deadline: 13 December 2026, 11:30 UTC
  • Results: 29 January 2027

Prizes and Benefits

  • 1st place: US$1,500, plus an invitation for one person to present the winning solution at the ChangeNOW Summit 2027 (worth US$2,500)
  • 2nd place: US$1,000
  • 3rd place: US$500
  • Best write-up: 2 × US$1,000
  • Experience with time-series, feature engineering and acoustic data on a real ocean conservation problem
  • A public leaderboard result and Zindi points to show employers

Eligibility

  • Open to all, as individuals or teams of up to four
  • One Zindi account per person; private code sharing outside your team is not allowed
  • Public external acoustic datasets may be used if you share their sources on the discussion board
  • Write-ups can come from any participant, whatever their leaderboard position

How to Apply

  1. Create a free account on Zindi.
  2. Open the Underwater Vessel Noise Challenge page, accept the rules and click Join.
  3. Download the data, build your model and submit predictions on the leaderboard.
  4. Optionally, submit a research write-up (PDF) and runnable code or notebook (10 MB maximum per file) through the form on the challenge page. Show pressure maps and compare against a simple baseline.
  5. Finish everything before 13 December 2026, 11:30 UTC.

Tips

  • Start with a simple baseline, such as energy in typical ship-noise frequency bands, and improve from there. The write-up judges want a baseline comparison.
  • If your model uses AIS data, also report results without it, as the rules ask.
  • Students with strong writing skills can target the US$1,000 write-up prizes even without a top leaderboard score.
  • Free tools such as Google Colab and Kaggle notebooks are enough for this tabular time-series data.
  • Remember that the deadline is 11:30 UTC, which is just after noon in Lagos, not midnight.

Related Opportunities

Leave a Reply

Your email address will not be published. Required fields are marked *