In this meeting, we gather together people who work on evaluation/benchmarking of machine learning for bioacoustic tasks – typically, evaluating systems that aim to recognise animal sounds. There are now many such systems, and there are also various evaluation campaigns. What should future bioacoustic evaluation initiatives be? How can we give the most benefit to our research communities and to the potential users of such systems?

Oct 30, 2024, 13:00–17:30 UTC
(Europe CET: 14:00-18:30)
(US West coast: 6:00-10:30)
(US East coast: 9:00-13:30)

Draft schedule:

(Note: times are given in UTC, also known as GMT or London time.)

  • 13:00 Dan Stowell (Naturalis / Tilburg University):
    “Benchmarks and evaluations: what do computational bioacoustic communities do, and what do they need?” (Including some background on prior DCASE bioacoustics)
  • 13:30 Existing evaluation campaigns & position statements:
  • 15:00 Break
  • 15:30 (TBC)
  • 16:00 Whiteboard session: bioacoustic tasks
  • 16:30 DCASE 2025 – Frederic Font Corbera (Universitat Pompeu Fabra)
  • 16:50 BioDCASE 2025 – Ester Vidaña-Vila (Ramon Llull University)
  • 17:00 Initial open planning discussions
  • 17:30 Close

To attend online, please email.

Written by

Dr. Dan Stowell

Read more about our research