BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY: Lessons from integrating Geospatial Foundation Models into enviro
 nmental journalism methodologies - Anne Alexander\, Cambridge Digital Huma
 nities
DTSTART:20251215T120000Z
DTEND:20251215T123000Z
UID:TALK241327@talks.cam.ac.uk
CONTACT:Sam Nallaperuma-Herzberg
DESCRIPTION:The use of foundation models produced from geospatial data\, s
 uch as satellite images\, is becoming increasingly common in ‘remote-sen
 sing’ methods deployed by journalists alongside traditional reporting me
 thods of field visits\, observation and interviews. This presentation disc
 usses the lessons learned from a project undertaken by CDH in collaboratio
 n with the Pulitzer Center and Watershed Investigations to deploy EarthInd
 ex\, a tool which combines code to produce cloud-free composite Sentinel-2
  images and run foundation model inference on 32x32-pixel patches worldwid
 e with an interface allowing users to label images and generate fast appro
 ximate nearest neighbour searches from a vector database. We were able to 
 test model predictions against investigation and observation at ground lev
 el and demonstrate the importance of triangulating information derived fro
 m the use of AI tools with other sources and perspectives. Our team member
 s’ field work also underlined the ethical challenges of relying on remot
 e sensing data and the potential for the use of such methods to harm vulne
 rable communities.
LOCATION:SS03 Seminar Room\, Willam Gates building (Department of Computer
  Science and Technology)
END:VEVENT
END:VCALENDAR