BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CorrelAid//Events//DE
NAME:Teaching Neural Networks to See Fine Urban Greenery
X-WR-CALNAME:Teaching Neural Networks to See Fine Urban Greenery
BEGIN:VTIMEZONE
TZID:Europe/Berlin
X-LIC-LOCATION:Europe/Berlin
LAST-MODIFIED:20260429T222214Z
BEGIN:DAYLIGHT
TZNAME:CEST
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
DTSTART:19700329T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZNAME:CET
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
DTSTART:19701025T030000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
TIMEZONE-ID:Europe/Berlin
X-WR-TIMEZONE:Europe/Berlin
BEGIN:VEVENT
UID:event-257@correlaid.org
SEQUENCE:0
DTSTAMP:20260616T010236
DTSTART;TZID=Europe/Berlin:20260702T190000
DTEND;TZID=Europe/Berlin:20260702T200000
SUMMARY:Teaching Neural Networks to See Fine Urban Greenery
LOCATION:Online
DESCRIPTION:We are the Cologne chapter of CorrelAid\, a network of volunte
 ers using their expertise on data for social good. Join us for our upcomin
 g event!\n\nThis talk focuses on the practical and scientific challenges o
 f detecting fine-scale urban vegetation from Sentinel-2 satellite imagery 
 (https://www.esa.int/Applications/Observing_the_Earth/Copernicus/Sentinel-
 2) under noisy and imperfect real-world conditions.\n\nSpecifically the fo
 llowing assumptions will be discussed:\n\n \n-     Assumption 1: AI ca
 n solve complex problems\n-    Assumption 2: Deep learning math is complic
 ated\n-      Assumption 3: BIG data is needed\n-     Assumption 4:
  Foundational models are generalizable\n-      Assumption 5: AI makes 
 Data Science superfluous\n\nHere is a publication on the topic for those w
 ho would like to read ahead:\n\n   -  EarthShift: a benchmark for meas
 uring robustness to real-world distribution shifts in Earth observation (h
 ttps://arxiv.org/abs/2605.29330)\n    - Code and data: https://earthsh
 ift.github.io\n\nOur speaker Dorothea is a PhD trained Biochemist who work
 ed in the intersect of Screening and Material Sciences with industry exper
 ience. She is a bit of a #Data4Good (https://correlaid.org/en/projects) en
 thusiast and try to make time for it. Here is a perspective piece which ca
 me out from a project Dorothea lead:\n\n  -   Blog Series on Learning 
 with Uncertain Multi-Band Images — Part 1: Why Mapping Urban Greenery Wa
 s Still Hard in 2025 (https://medium.com/@edp_2023/blog-series-on-learning
 -with-noisy-multi-band-images-part-1-why-mapping-urban-greenery-was-725113
 78a3c0)\n\nFor those who are new to CorrelAid: you are very welcome to joi
 n! We can always chat 1:1 after the main event on all you'd like to know a
 bout CorrelAid and how to participate in our volunteer projects to build u
 p your portfolio while helping others.\n\nPlease note the following:\n\n 
 This event follows the CorrelAid Code of Conduct (https://correlaid.org/en
 /coc)   \n\n This event will be conducted in English.\n \n There will be N
 O recording of the session.\n \n AI bots for transcription are not allowed
 .\n \n There are multiple channels to enroll in this event. As such the nu
 mber of persons signing up on Meetup does not reflect the eventual number 
 of persons participating.\n\nRegistration: https://www.meetup.com/correlai
 d-cologne/events/315137621/\n\nView more at: https://correlaid.org/veranst
 altungen/lc-cologne-jul-26
URL;VALUE=URI:https://correlaid.org/veranstaltungen/lc-cologne-jul-26
END:VEVENT
END:VCALENDAR