BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Barlow Twins Foundation Model (BTFM) for Earth Observation - Speak
 er to be confirmed
DTSTART:20250520T130000Z
DTEND:20250520T140000Z
UID:TALK231871@talks.cam.ac.uk
CONTACT:Yi Zhang
DESCRIPTION:Satellite imagery provides a critical lens for monitoring Eart
 h’s dynamic systems\, yet integrating multi-source\, multi-temporal data
  into globally consistent\, high-resolution representations remains a chal
 lenge. Traditional remote sensing vision models\, which process patches or
  images as inputs\, often struggle to capture fine-grained spatiotemporal-
 spectral relationships critical for downstream tasks like land classificat
 ion\, climate modeling\, and change detection. We present a self-supervise
 d framework leveraging Barlow Twins to train an Earth Foundation Model tha
 t outputs pixel-level representations from diverse satellite data sources.
  Preliminary results demonstrate that the resulting representation map enc
 odes high-quality spatiotemporal patterns\, outperforming other earth foun
 dation models (including current SOTA model\, Google’s Embedding Fields 
 Model) in classification and regression tasks. By bridging multi-modal sat
 ellite data into a harmonized latent space\, our approach unlocks new oppo
 rtunities for monitoring planetary-scale processes with higher precision.\
 n
LOCATION:Main Seminar Room (1.25)\, David Attenborough Building
END:VEVENT
END:VCALENDAR