BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Using SSL to create useful representations from corrupted spectral
 -temporal data for applications in agriculture - Madeline Lisaius - EEG
DTSTART:20230303T130000Z
DTEND:20230303T140000Z
UID:TALK197641@talks.cam.ac.uk
CONTACT:Madeline Lisaius
DESCRIPTION:Agriculture is central to questions of poverty\, climate chang
 e\, land use\, and human and planetary health. Managing agricultural lands
 capes at regional and global scales necessitates equally large-scale data:
  satellite remote sensing imagery is a natural fit. Current methods in rem
 ote sensing have historically and are currently challenged by corrupted an
 d missing data as well as the question of how to use the vast volumes of e
 xisting unlabeled data. We propose a new method and new data construct –
  an adaptation of Barlow Twins using the “d-pixel” – which we postul
 ate will be effective for downstream tasks while both invariant to missing
  data and leveraging unlabeled data. 
LOCATION:Seminar time is 1pm BST in Room FW 11\, Willam Gates Hall. It wil
 l only be hosted in person.
END:VEVENT
END:VCALENDAR