The #DCASE2021 evaluation sets are now released! Including our few-shot bioacoustic event detection: http://dcase.community/challenge2021/task-few-shot-bioacoustic-event-detection Good luck to everyone who is working on #machinelistening...
Nice piece of work: research from @_BTO has investigated the effect of artificial light at night on birds (via acoustic monitoring and some deep learning) https://bto.org/our-science/publications/peer-reviewed-papers/nocturnal-flight-calling-behaviour-thrushes-relation
Aha, a new strongly-annotated dataset of birdsong audio! (from Eastern USA) https://esajournals.onlinelibrary.wiley.com/doi/full/10.1002/ecy.3329 (thanks @whybirdssing for the tip) #bioacoustics
1st International Workshop on Automatic Translation for Sign and Verbal Languages https://sites.google.com/tilburguniversity.edu/at4svl2021/home
One thing common across projects: data "de-duplication" / dataset "conflation" algorithms. Most of us hand-roll a method for our own situation. Imagine a nice Python library to merge+deduplicate an arbitrary set of energy-system geodata...
Looking forward to taking part in the @openmod lightning talks about "open global datasets for energy system modeling" today at 15:00 CEST: https://forum.openmod.org/t/4rd-online-lightning-talk-mini-workshop/2574
Leiden (NL) dawn chorus this morning - we heard: Jay, Wren, Chicken, Great spotted woodpecker, Mallard, Magpie, Great tit, Blue tit, Wood pigeon, Chiff chaff, Long-tailed tit, Chaffinch, Goose, Song thrush, Parakeet, Robin, Blackbird, Gull, Carrion crow, Greenfinch. #Reveil2021
The dawn has now reached Japan. We're watching (remotely) "Cyberforest" who are conducting a birdsong census in a Japanese forest, as part of the @soundtent Reveil http://soundtent.org/reveil/
At the Dutch seaside today, we heard SO many nightingales! Fascinating sound to hear close up. Here's a clip recorded using @warblr:
Everyone should join the Soundcamp Reveil this weekend http://soundtent.org/ -- a beautiful celebration of dawn chorus birdsong beamed from around the world, with added sonic art
A very nice summary of "Google's tips for reducing the CO2 emissions of training AI models" by @layon_overwhale: http://revue.dynamicallytyped.com/issues/64-google-s-tips-for-reducing-the-co2-emissions-of-training-ai-models-555894 #deeplearning
Can open data help to fix the energy system's CO2 impact? I'm looking fwd to taking part in this set of lightning talks on 5th May - check out the list of topics: https://forum.openmod.org/t/4rd-online-lightning-talk-mini-workshop/2574/2
Server run by the main developers of the project It is not focused on any particular niche interest - everyone is welcome as long as you follow our code of conduct!