The flock is now being asked to file its own flight plan
the flock is now being asked to file its own flight plan: a low cost tracker detects and follows birds on x band marine radar at a tidal test site, self supervised learning pulls waterbird movement out of weather radar without hand labels, and a kaggle challenge now trains models to classify species from wind farm radar tracks.
the sky now files its own migration log.
Context
Autonomous Radar Ornithology (May 9, 2026) evaluates GANNET, a low-cost algorithm for bird-like targets in X-band marine radar imagery at the Fall of Warness tidal test site in Orkney. The study reports more than 55,000 filtered detections and 34,000 trajectories, and processing over ten times faster than manual annotation.
Detection of Animal Movement from Weather Radar using Self-Supervised Learning (August 8, 2024) pretrains on noisy labels produced by thresholding, then fine-tunes on a small human-labelled dataset. Its Australian waterbird experiment reports a 43.53% improvement in Dice coefficient relative to its comparison method. AI Cup 2026’s Kaggle performance track asks models to classify bird species from radar tracks collected at windfarm Eemshaven in Groningen; that competition description was available in search excerpts, not in the fetched page.
Self-supervised does not mean no hand labels anywhere: this method explicitly fine-tunes on human labels. GANNET tracks bird-like targets rather than identifying species, and achieves about 80% recall and 70% precision in good-quality imagery, with both degrading under stronger clutter. The Kaggle leaderboard covers only the performance track, not the entire AI Cup. Its description remains excerpt-only here.
Watch next
- Species-level errors and detection of small birds under sea clutter. The labelled fine-tuning set size behind the waterbird result.
Sources
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 4 October 2026 at 08:08 IST. Sources are the papers and datasets the note draws on.
View the original post ↗Embed this note
More notes
The air is now being asked to keep its own ledger
the air is now being asked to keep its own ledger: ecmwf’s aifs compo becomes the first ai model to forecast atmospheric composition globally every three hours, cleanair simulates 365 days of pm2.5 over china in ten seconds, and a unified framework maps six pollutants at one kilometer across the whole country. the air now files its own composition report.
read the note →The current is now being asked to draw its own map
the current is now being asked to draw its own map: china’s langya 2.0 predicts six ocean phenomena including internal waves and mesoscale eddies, a deep net called wenhai resolves eddies globally with air sea flux formulas built in, and scripps infers surface currents from the way temperature patterns deform in satellite images. the ocean now files its own circulation report.
read the note →The soil is now being asked to report its own carbon
the soil is now being asked to report its own carbon: a nix color sensor paired with generative data augmentation predicts soil organic carbon without a lab, random forest drives 74 percent of soil health mapping studies, and sentinel 2 tracks five year carbon change across france and italy from 922 samples. the dirt now files its own carbon account.
read the note →