Sign in

Callan Alexander

@callanalexander.bsky.social
420 followers 273 following 28 posts

PhD Candidate at QUT, Threatened Species Technical Coordinator at BirdLife Australia. Bioacoustics & Machine Learning Music: www.clnmusic.com/about

PostsRepliesMedia
Callan Alexander @callanalexander.bsky.social · 17/09/2026
I've found this to be really useful in my work validating bird detections. It's saved me hours of sifting through false-positives and helped me to find vocalisations for difficult species that I otherwise would have missed. 🦉 A big thanks to co-authors Ayesha Tulloch, Paul Roe, and Susan Fuller.
000
Callan Alexander @callanalexander.bsky.social · 17/09/2026
We tested this on several species and about 65,000 potential detections. The clustering workflow removed almost all of the false positives (~98%) in our dataset & retained more true detections than any threshold from the supervised models.
100
Callan Alexander @callanalexander.bsky.social · 17/09/2026
We are also releasing pardalote, an open-source tool that allows users to iteratively cluster sounds by acoustic similarity (using deep-learning embeddings), easily delete unwanted 'noise' and re-cluster/keep & label the remaining audio.
110
Callan Alexander @callanalexander.bsky.social · 17/09/2026
We found that you can use an iterative clustering process to locate vocalisations across the entire threshold range, and easily discard false-positives and keep true detections.
100
Callan Alexander @callanalexander.bsky.social · 17/09/2026
Setting confidence thresholds for supervised audio models (like BirdNET or Perch) is often a somewhat frustrating trade-off between precision and recall. You are trying to balance finding as many vocalisations as possible while minimising the amount of false-positives that you need to sift through.
100
Callan Alexander @callanalexander.bsky.social · 17/09/2026
New preprint! In this paper, currently in review (lnkd.in/gGGvSktM), we demonstrate a hybrid human-in-the-loop machine learning workflow for validating bioacoustic classifier outputs. #bioacoustics #machinelearning #acousticmonitoring #BirdNET #Perch
260
Callan Alexander @callanalexander.bsky.social · 28/11/2025
Just finished up a big few weeks with four talks at three conferences. Thanks for having me #Acoustics2025, #AOC2025 and #ESA2025. So good to learn from so many brilliant people. Also managed to find some time to see some wild Numbats + a few other cool plants & animals. Now time for a nap 😴
0140
Callan Alexander @callanalexander.bsky.social · 13/09/2025
thanks @ecologygrant.bsky.social! Sorry I missed this post, have been in the depths of PhD write up and neglected my Bluesky account.
130
Callan Alexander @callanalexander.bsky.social · 09/06/2025
I’m hoping to look at translation to other species soon. There are some tricky bits when dealing with overlapping calls and complex song types. I think that there are ways to get around that though, but how well it performs ultimately may vary depending on the species and region. 🤷
000
Callan Alexander @callanalexander.bsky.social · 09/06/2025
I guess at this stage I mostly mean that you can apply this to huge acoustic datasets and collect note annotations at very large ‘scales’ without requiring much more effort. That being said, I think it will work well on other species with some tweaking!
100
Callan Alexander @callanalexander.bsky.social · 02/06/2025
This approach is highly scaleable, and can be used to rapidly 'harvest' individual note annotations from large acoustic datasets. This allows for further investigation of song structure, geographic call variation and potentially vocal individuality.
000
Callan Alexander @callanalexander.bsky.social · 02/06/2025
The clustering process is also very effective at searching 'underneath' a classifier threshold, removing false positives and finding vocalisations that may have been missed. The red circles in the figure show where the owl vocalisations have been split away from other sounds.
100
Callan Alexander @callanalexander.bsky.social · 02/06/2025
In this paper we use a multi-stage machine learning pipeling (combining supervised and unsupervised methods) to reduce almost 3000 hours of environmental recordings into 10,116 annotations, of which 93% were correctly annotated individual notes of the target species.
130
Callan Alexander @callanalexander.bsky.social · 02/06/2025
New paper out now in Ecological Informatics! www.sciencedirect.com/science/arti... #bioacoustics #machinelearning #acousticmonitoring
sciencedirect.com
Automated note annotation after bioacoustic classification: Unsupervised clustering of extracted acoustic features improves detection of a cryptic owl
Passive acoustic monitoring and machine learning are increasingly being used to survey threatened species. When automated detection models are applied…
2133
Callan Alexander @callanalexander.bsky.social · 19/05/2025
Wasn’t expecting to see this on my walk yesterday - Noisy Pitta in suburban Brisbane! They’re usually rainforest birds but migrate at this time of year and can show up in weird places. Haven’t added any new species to my home patch list (now at 98) in ages, this will be hard to beat.
030
Reposted by Callan Alexander
David Marler @qldaah.bsky.social · 03/05/2025
It's Qld Labor in Bonner. This is massive. This seat hasn't moved from Qld LNP since 2010. #ausvotes
27212
Reposted by Callan Alexander
Martine Maron @martinemaron.bsky.social · 14/02/2025
Birding is booming in Australia @birdlifeoz.bsky.social www.abc.net.au/news/2025-02...
abc.net.au
How birdwatching is taking flight as Australia's new tourism goldmine
Birdwatchers are pouring billions of dollars into Australia's tourism industry every year with calls to capitalise on the country's native species and better promote the pastime.
0155
Reposted by Callan Alexander
scottwilliamf.bsky.social @scottwilliamf.bsky.social · 03/02/2025
Stoked to have a cover photo for @ecography.bsky.social! In our paper doi.org/10.1111/ecog... we included temporal dynamics into SSFs, resulting in daily patterns of movement and habitat selection. Simulating gave us dynamic spatial predictions across the landscape. Code! github.com/swforrest/dy...
1154
Reposted by Callan Alexander
Open Ecoacoustics @openecoacoustics.bsky.social · 31/01/2025
Introducing Open Ecoacoustics: a platform for continental-scale ecological monitoring and research in Australia. As part of ARDC's Machine Observation Data Processing Infrastructure, we use ecoacoustic technologies to promote open science and conservation. #OpenScience #Ecoacoustics #bioacoustics
doi.org
Open Ecoacoustics | ARDC
The Open Ecoacoustics platform is enabling continental-scale ecological monitoring and research.
1164
Reposted by Callan Alexander
Charlotte Patterson @charlotterubyp.bsky.social · 30/01/2025
Excited to present our latest paper where we take a close look at what we know (and are yet to know) about biodiversity in Antarctica! 🔗 Check out a summary of the work: arcsaef.com/story/new-an... 🔗 Read the paper here: doi.org/10.1111/ddi....
24914
Reposted by Callan Alexander
Grant Hamilton 🐨 @ecologygrant.bsky.social · 24/01/2025
Are you a scientist that analyses/creates models using data from cameras, acoustics, drones or satellites to understand biodiversity? Join the Biodiversity Monitoring starter pack. Point me to your research to be added. ECRs and PhDs welcome! 🧪🌏 #AI #conservation go.bsky.app/eFMQQX
238230
Callan Alexander @callanalexander.bsky.social · 23/01/2025
Thank you! Cool feature, would be keen to be on the contributors list!
110
Callan Alexander @callanalexander.bsky.social · 21/01/2025
The pre-print is still a bit rough, so very open to any feedback during the review process! The next step is to try this on other species and calls. It may be a bit trickier with diurnal species or those with more complex vocalisations.
030
Callan Alexander @callanalexander.bsky.social · 21/01/2025
Anyone who works in bioacoustics knows that tagging individual notes is extremely time consuming, so I think this is a really good way to 'harvest' note annotations from noisy environmental data. You can then use these annotations to compare geographic call variation or even vocal individuality.
140
Callan Alexander @callanalexander.bsky.social · 21/01/2025
We then found that applying iterative unsupervised clustering (using UMAP and HDBSCAN) to the acoustic features was useful for separating false-positive and true detections, and in the end you are also left with individual note annotations that have essentially been automatically tagged!
100
Callan Alexander @callanalexander.bsky.social · 21/01/2025
In this paper (currently in review) we apply a hybrid approach to automated detection of a cryptic threatened owl, the Powerful Owl (Ninox strenua). Step one is a neural network to classify the vocalisations, after which we segment the output and extract acoustic features.
100
Callan Alexander @callanalexander.bsky.social · 21/01/2025
New pre-print! 🎙️🦉 Automated Note Annotation after Bioacoustic Classification: Unsupervised Clustering of Extracted Acoustic Features Improves Detection of a Cryptic Owl papers.ssrn.com/sol3/papers.... #bioacoustics #machinelearning #conservation
34211
Callan Alexander @callanalexander.bsky.social · 30/12/2024
First time birding in NZ. Cannot believe how many exotics there are everywhere. House sparrow chatter definitely most common noise of the day (Blackbird song close second). Loved the Gannet colony!
070
Callan Alexander @callanalexander.bsky.social · 12/12/2024
Thanks so much for watching!
100
Reposted by Callan Alexander
Alex Slavenko @alexsla.bsky.social · 12/12/2024
@callanalexander.bsky.social on birds in vineyards - heaps of native birds present, and using a cool combination of supervised and unsupervised machine learning models to overcome false positive issues in AI detection of bird calls #ESAus2024
2131
Reposted by Callan Alexander
Dr. Hilary Rose Dawson @hilaryrosed.bsky.social · 08/12/2024
Good morning, #ESAus2024! Lots of people joined the starter pack yesterday. Take a(nother) look for other austral ecologists to follow and check out the hashtag for interesting talks coming up today. go.bsky.app/5YmZLNm
0269
Callan Alexander @callanalexander.bsky.social · 08/12/2024
Will be there, list is a great idea!
110
Callan Alexander @callanalexander.bsky.social · 08/12/2024
Excited for #ESAus2024 this week! I'm speaking about birds, wine & machine learning on Thursday afternoon.
081
Callan Alexander @callanalexander.bsky.social · 08/12/2024
Would like to be added if possible please! I work on AI + threatened species monitoring in Australia. :)
020
Callan Alexander @callanalexander.bsky.social · 08/12/2024
Would love to be added please! Great list.
110
Reposted by Callan Alexander
Thejasvi Beleyur @beleyurthejasvi.bsky.social · 20/11/2024
Hello Bioacousticians, Turns out there wasn't a bioacoustics starter pack. Just realised I myself am following the bare minimum number required to make a starter pack. Please share & nominate yourself! Hoping to see this pack grow. go.bsky.app/2Qmjek6
24179
Callan Alexander @callanalexander.bsky.social · 07/12/2024
Finally joined BlueSky! I'm a PhD candidate mostly working on #bioacoustics, #machinelearning & #conservation. This black squiggle (a spectrogram) may not seem like much, but it’s the first Eastern Bristlebird heard in QLD in 5+ years found using acoustic recorders + a deep learning model.
2213