BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Calibrated Probabilistic Interpolation for GEDI Biomass - Robin Yo
 ung\, University of Cambridge
DTSTART:20260123T130000Z
DTEND:20260123T140000Z
UID:TALK241927@talks.cam.ac.uk
CONTACT:114742
DESCRIPTION:*Abstract*\nMapping global forest biomass from NASA's GEDI mis
 sion requires interpolating sparse LiDAR observations across diverse lands
 capes. Standard machine learning approaches like Random Forest and XGBoost
  fail to produce calibrated uncertainty estimates\, as they conflate ensem
 ble variance with true predictive uncertainty and ignore spatial context.\
 nWe introduce Attentive Neural Processes (ANPs)\, a probabilistic meta-lea
 rning framework that conditions predictions on local observations and geos
 patial foundation model embeddings. ANPs learn flexible spatial covariance
  functions\, expanding uncertainty in complex landscapes and contracting i
 t in homogeneous areas. Validated across five biomes from tropical Amazoni
 an to boreal and alpine ecosystems\, ANPs achieve competitive accuracy wit
 h near-ideal uncertainty calibration. The framework also enables few-shot 
 adaptation\, recovering most cross-region transfer performance with minima
 l local data. This provides a scalable\, principled alternative to ensembl
 e methods for continental-scale biomass mapping.\n\n*Bio*\n\nRobin Young i
 s a first-year PhD student in Computer Science at the University of Cambri
 dge.
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