BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//pretalx.earthmonitor.org//global-workshop-2026//9PTJVH
BEGIN:VTIMEZONE
TZID:Europe/Amsterdam
BEGIN:STANDARD
DTSTART:20001029T030000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10
TZNAME:CET
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20000326T020000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3
TZNAME:CEST
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
UID:pretalx-global-workshop-2026-3MAMRS@pretalx.earthmonitor.org
DTSTART;TZID=Europe/Amsterdam:20261009T113000
DTEND;TZID=Europe/Amsterdam:20261009T120000
DESCRIPTION:Deep learning represents a powerful tool to interpret Earth obs
 ervation data at large geographic scales. However\, in cases where abundan
 t reference data is not available and cannot easily be collected\, new app
 roaches are needed to benefit from this technology. Several Earth observat
 ion tasks\, especially in environmental remote sensing\, remain challengin
 g due to the limited number of samples and the geographic and temporal bia
 s in the reference data. Furthermore\, mapping biophysical variables from 
 single sensor inputs often leads to high ambiguities. Multimodal models pr
 etrained in a self-supervised fashion promise to overcome such challenges.
 \n\nIn this talk\, I will first present our recent research project MMEart
 h-Bench\, a multimodal benchmark dataset for environmental remote sensing.
  I will discuss our evaluation of existing pretrained models and present o
 ur test-time adaptation approach that adapts any model at test time using 
 multimodal data to construct adaptation signals. Lastly\, I will present S
 uperF\, an approach for multi-image super-resolution. This test-time optim
 ization approach based on implicit neural representations makes use of rep
 eated observations with sub-pixel shifts and does not require any high-res
 olution training data. Under a static scene assumption\, it can be applied
  to super-resolve e.g. Sentinel-2 time series for any place on Earth.\n\n-
  Personal website: https://langnico.github.io/\n- MMEarth-Bench project: h
 ttps://mmearth-bench.com/\n- SuperF project: https://sjyhne.github.io/supe
 rf/
DTSTAMP:20260624T071306Z
LOCATION:Aula Magna
SUMMARY:Optimizing Representations at Test Time - Nico Lang
URL:https://pretalx.earthmonitor.org/global-workshop-2026/talk/3MAMRS/
END:VEVENT
END:VCALENDAR