Ioannis Prapas

🔥Wildfire Danger Prediction and Understanding with Deep Learning Deep Learning models to predict Wildfires and xAI to understand what they have learned. 📜Paper: doi.org/10.1029/2022GL… accepted in @theAGU GRL 🧊Datacube: zenodo.org/record/6475592 💻Code: github.com/Orion-AI-Lab/w…

🚨 Live at #LPS25 in Room 0.11/0.12! NTUA & TUM in action, leading a tutorial on Foundation Models (FM) for EO — practical AI for land cover mapping, object detection & more 🌍🛰️ Interested in FM?👉 thinking-earth.eu/foundation-mod… @ThinkingEarthEU @NikosGiannisB #RemoteSensing #AI

Deep search and satellite data is an interesting mix 🔥🛰️ The demo shows the AI Reporter doing deep investigation into the Palisades fire early in 2025. In ~2 minutes you get a concise report, based on deep search and satellite data. #geoAI #AI4EO #PalisadesFire #AgenticAI

🚀 IGARSS kicks off this Sunday, and our lab will be there with an amazing tutorial! 🧠💻 Join @yannispapoutsis, @iprapas, @skondylatos, @NikosGiannisB, and Maria for a deep dive into AI and natural hazards. Info below⬇️ #IGARSS2024 #AI #Wildfires #Floods #DataCubes

Join the Scaling Science workshop at #IGARSS2024 to master scaling geospatial data & scientific research w/ open-source tools! Ideal for scientists, researchers, & practitioners looking to leverage cloud computing, machine learning, & big data analytics. #tiescaling" target="_blank" rel="nofollow noopener">2024.ieeeigarss.org/tie_events.php…

@vitusbenson @EarthNetTeam @ClaireRobin01 @crequena2 @LazarusAlon @Zhihan_Gao @Reichstein_BGC @Ellis_Unit_Jena Great work, congratulations!

@EarthNetTeam Pre-Print: arxiv.org/abs/2303.16198… In collaboration w/ many: @ClaireRobin01 @crequena2 @LazarusAlon @Zhihan_Gao @Reichstein_BGC @Ellis_Unit_Jena

Our work from @EarthNetTeam has been accepted to CVPR 😊 We benchmark many vegetation forecasting models on the improved GreenEarthNet dataset + introduce the transformer-based Contextformer. This gif shows predictions for various weather scenarios - towards AI 4 Early Warning ✨

Excited to share that SPOT is accepted to #CVPR2024. SPOT advances unsupervised object-centric learning with attention-based self-training & patch-order permutation, achieving state-of-the-art results. - Paper: arxiv.org/abs/2312.00648 - Code: github.com/gkakogeorgiou/…

@skondylatos is in New Orleans @neuripsconf, presenting the Mesogeos dataset, a large-scale dataset for wildfire modeling in the Mediterranean. Don't miss his oral presentation at 15.30 local time #neurips orion-ai-lab.github.io/publication/ko… #neurips

🌍 🚨📜🤖👨‍👩‍👧‍👧 Machine learning helps us unravelling the climate interactions between socioeconomic context and natural hazards on human population displacement We study The Displacement Puzzle with #AI! Read it @NatureComms: nature.com/articles/s4146… Thread🔄

[2/2] ✍ Authors: Nikolaos Ioannis Bountos @NikosGiannisB, Ioannis Papoutsis @yannispapoutsis, Dimitrios Michail @TheDMichail and Nantheera Anantrasirichai 🔗DOI: doi.org/10.1109/LGRS.2… #earthobservation #remotesensing #ai #Volcanoeruption #volcano

[1/2] Learn how #SAR plays a crucial role in identifying #Fagradalsfjall volcanic unrest in #Iceland ! 🆕 GRSS article of the week 📕Journal: #IEEE Geoscience and Remote Sensing Letters 📑Title: Self-Supervised Contrastive Learning for Volcanic Unrest Detection

TeleViT "Teleconnection-driven Transformers Improve Subseasonal to Seasonal Wildfire Forecasting" won the Best Paper Award 🏆 at the Artificial Intelligence for Humanitarian Assistance and Disaster Response Workshop of @ICCVConference #ICCV2023 🇫🇷

Orion Lab's @iprapas at @ICCVConference in Paris 🇫🇷 🗓️Spotlight Talk today at 15:45 CET + Poster Session 👉 TeleViT: Teleconnection-driven Transformers Improve Subseasonal to Seasonal Wildfire Forecasting #ICCV2023 hadr.ai/iccv23/iccv-sc…

Presenting tomorrow 3/10 at 15:45 TeleViT in AI+HADR #ICCV2023 @ICCVConference Paris. Join us also for the poster session right after. Full schedule: hadr.ai/iccv23/iccv-sc…

📢 Join DeepCube, @CALLISTO_H2020 and @EnvisionH2020 at the NASA International Space Apps Challenge in Thessaloniki! spaceappschallenge.org DeepCube will participate in the #hackathon with a challenge on #wildfire danger forecasting 👩‍💻👨‍💻 More details coming soon!

This is an output of the ESA-funded project SeasFire github.com/seasfire. The project is implemented by @OrionLab_NOA , @harokopio_ and @MPI_BGC . #esa #philab @blesa_ux

Great collaboration with @NikosGiannisB , @skondylatos, Dimitrios Michail, @isp_uv_es, @yannispapoutsis

🧵 (1/6) 🔥 "TeleViT: Teleconnection-driven Transformers for Subseasonal to Seasonal Wildfire Forecasting" to be presented at the ICCV 2023 AI+HADR workshop. 💻Code github.com/Orion-AI-Lab/t… 🌐Website orion-ai-lab.github.io/televit 📜Paper arxiv.org/abs/2306.10940