Lily-belle Sweet
PhD student at UFZ - interested in explainable machine learning, agriculture and food security, compound climate events 🌾
- I'm really looking forward to leading a student research project at the ELLIS Summer School on AI for Earth and Climate Sciences in Jena this September! What a great lineup of speakers, mentors and tutors www.ellis-jena.eu/summer-schoo... @ellisunitjena.bsky.social @ellis.eu
- So excited to be organising this. Don't forget to register if you'd like to attend! We want to bring together people from different disciplines, so if your background is ML without the Ag, or vice versa, but you're interested in learning a bit about the other side - join us in Leipzig (or online)
- 📢 The 2nd AgML Workshop will be held at @ufz.de Leipzig from November 3–5th! Let's bring together researchers and practitioners in agriculture and machine learning to share knowledge & create open benchmark datasets and model evaluation protocols. Register by July 31: events.hifis.net/e/agml2025
- 📢 The 2nd AgML Workshop will be held at @ufz.de Leipzig from November 3–5th! Let's bring together researchers and practitioners in agriculture and machine learning to share knowledge & create open benchmark datasets and model evaluation protocols. Register by July 31: events.hifis.net/e/agml2025
- 📢 Tomorrow (27th May) at 16:00 CET, join our next online seminar on knowledge-guided machine learning for crop growth modelling! We will be hearing from Jingye Han, a researcher at Wageningen University. Open to all - get in touch for the Zoom link or join the AgML mailing list.
- If you, like us, think that these kinds of community activity are needed, we warmly invite you to get involved in AgML. So far, we've created benchmark datasets, organised a workshop, Kaggle competition and there's so much more we'd like to do - join our mailing list: mail.agml.org/mailman/list...
- The peat map issues are in the Times today (@catframpton.bsky.social and I are quoted) - www.thetimes.com/uk/environme...
- Looked at the report to see how they evaluated model performance and what a surprise - the reported accuracy metrics are based on randomly sampled datapoints. Spatial data is not i.i.d., and model evaluation is the hardest part of any ML project (but not as fun as trying a dozen model architectures)
- More and more peat map issues identified. If only all ML maps used in academic research got the kind of painstaking sanity-checking that Cat's been doing
- Reposted by Lily-belle Sweet[Not loaded yet]
- Looked at the report to see how they evaluated model performance and what a surprise - the reported accuracy metrics are based on randomly sampled datapoints. Spatial data is not i.i.d., and model evaluation is the hardest part of any ML project (but not as fun as trying a dozen model architectures)
- For vegetation modelling they use XGBoost, and it's hard to parse the exact method but it sounds like they use either random or stratified random sampling to define their test set (either of which would be incorrect for this application)...
- Reposted by Lily-belle Sweet🔥🌍 In a new paper led by Sifang Feng, we find a growing human-induced climate change fingerprint in weekly regional fire extremes. 📈 On average, climate change was responsible for a fraction equal to 8 ± 4% of the predicted probability of more than 700 fire extremes worldwide between 2002 and 2015.
- 📢 Today at 16:00 CET: Vasileios Sitokonstantinou presents research on invariant (causal) learning for robust yield predictions, kicking off our monthly seminar series on AI for agricultural research & modelling. Reach out for Zoom link, mailing list members will receive it by email. See you there!
- If you, like us, think that these kinds of community activity are needed, we warmly invite you to get involved in AgML. So far, we've created benchmark datasets, organised a workshop, Kaggle competition and there's so much more we'd like to do - join our mailing list: mail.agml.org/mailman/list...
- Transdisciplinary coordination is essential for advancing agricultural modelling with machine learning. The opportunities are great, but to avoid pitfalls, we need to establish benchmarks, evaluation criteria and best practices. Our perspective: www.cell.com/one-earth/fu... @cp-oneearth.bsky.social
- Written together with @inathens.bsky.social, @ronvbree.bsky.social, Andres Castellano, Pierre Martre, Dilli Paudel, @alexruane.bsky.social, @zscheischlerjak.bsky.social. Thank you to everyone involved in AgML and @agmip.bsky.social - the enthusiasm and passion of this community is so special
- Reposted by Lily-belle SweetWe are hiring 3 PhDs at the UFZ PhD Cohort SEESAW - Societal and environmental impacts of complex extremes in a changing world I'll be supervising a project looking at the interactions of flood and drought impacts 💧🔆 Apply here: recruitingapp-5128.de.umantis.com/Vacancies/31... #academicjobs
- Another nice office for the morning - the State Library of Western Australia in Perth
- Reposted by Lily-belle Sweet[Not loaded yet]
- Another day, another office...the Edge at State Library Queensland. Great view, lovely coffee!
- Reposted by Lily-belle Sweet[Not loaded yet]
- Saying goodbye to Gold Coast & its beautiful orb-weaver spiders - I'll be in Brisbane until Saturday (visiting CSIRO on Friday & hopefully giving a short talk). Get in touch if you're in the area!
- Reposted by Lily-belle SweetExcited to share our recent paper led by Sophie Johnson on the relationship between sequential humid heat and precipitation events 🌡️☔ We identify where humid heat is associated with suppressed or enhanced rainfall, with implications for compound extremes! iopscience.iop.org/article/10.1...
- Reposted by Lily-belle SweetIf you're new to #Bluesky, our #Plant Modelling Scientists starter pack can help you find users you can follow to help you get started: bsky.app/starter-pack... #PlantScience Let us know in the comments or DM if you'd like to be added.at://did:plc:oqqpxzlqy7m7z2zqps3rjrts/app.bsky.graph.starterpack/3lbdjhpwhwa23
- Reposted by Lily-belle SweetComing soon: I’ll be hiring a PhD candidate to explore regions in Germany disproportionately affected by floods & droughts, focusing on cross-sector impacts and using NLP-driven datasets #academicjobs
- Reposted by Lily-belle Sweet[Not loaded yet]
- Reposted by Lily-belle Sweet[Not loaded yet]
- Welcome to my office (for today)...