BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Predicting Global Patterns of Mycorrhizal Fungal Biodiversity with
  Self-Supervised Satellite Features - Robin Young\, University of Cambridg
 e
DTSTART:20250710T120000Z
DTEND:20250710T130000Z
UID:TALK234145@talks.cam.ac.uk
CONTACT:114742
DESCRIPTION:*Abstract*\n\nSoil fungal communities are critical drivers of 
 terrestrial ecosystem function\, yet their global distribution remains lar
 gely unknown due to the challenges of widespread physical sampling. We dev
 eloped a machine learning pipeline to predict fungal biodiversity across E
 urope and Asia using high-resolution\, temporal satellite imagery. We intr
 oduce a novel feature set derived from a self-supervised learning (SSL) mo
 del applied to Sentinel time series. We trained a model on roughly 12\,000
  mycorrhizal fungal richness samples\, comparing the predictive power of o
 ur SSL features against standard environmental datasets. Our combined mode
 l achieves a robust R2 of 0.53-0.55 across 50 cross-validation runs. We sh
 ow that the SSL features are the single most important predictor group\, o
 utperforming traditional datasets and implicitly capturing land cover info
 rmation. Furthermore\, we demonstrate that prediction errors are geographi
 cally clustered in sparsely sampled regions\, providing a data-driven meth
 od for identifying "biodiversity data deserts" and guiding future sampling
  efforts. This work presents a scalable framework for monitoring an overlo
 oked component of global biodiversity and demonstrates the viability of te
 mporally-rich\, self-supervised representations for ecological modeling.\n
 \n*Bio*\n\nRobin Young is a first-year PhD student in Computer Science at 
 the University of Cambridge.
LOCATION:Room GS15 at the William Gates Building and on Zoom: https://cl-c
 am-ac-uk.zoom.us/j/4361570789?pwd=Nkl2T3ZLaTZwRm05bzRTOUUxY3Q4QT09&amp\;fr
 om=addon 
END:VEVENT
END:VCALENDAR