AI4CS focuses on advancing AI methods to model and understand complex systems, including the visual brain, Earth systems, and socio-environmental interactions. With growing observational data, improved mechanistic models, and advanced machine learning, we can now monitor, predict, and analyze diverse variables. However, current approaches face key challenges: inefficiency with complex data, lack of adherence to physical laws, and limited fundamental insights. AI4CS addresses these issues with innovative, physics-aware, causality-driven AI solutions. This workshop brings together researchers to share advancements, generate new ideas, and promote interdisciplinary collaboration to enhance understanding of interconnected systems and tackle pressing global challenges.

Venue: ADEIT (room 1.1 & 1.2)

Day 1 – Monday, 13th January

TimeSessionPresenterDetails
2:00 PM – 2:30 PMIntroduction to AI4CS and workshop goalsGustau Camps-VallsOverview of AI4CS, objectives, and key challenges.
Explanation of workshop structure, session formats, and expected outcomes.
2:30 PM – 4:00 PMSession 1 (S1) – Earth-Land SystemsMaría PilesGlobal soil moisture and microwave vegetation parameters for ecology
Maurizio MencucciniOpportunities for AI in forest ecology
David ChaparroAn overview of vegetation moisture retrievals from remote sensing
4:00 PM – 4:30 PMCoffee Break
4:30 PM – 5:30 PMSession 2 (S2) – Earth-Water SystemsAna Belén RuescasChallenges of AI for in-land and coastal water management
Álvaro Moreno-MartínezLeveraging Crowd-sourced Biodiversity Data for an Enhanced Plant Functional Trait Mapping
5:30 PM – 6:00 PMSession 3 (S3) – Earth-Atmosphere SystemsCosmin Madalin MarinaFeature Selection for extreme atmospheric events

Day 2 – Tuesday, 14th January

TimeSessionPresenterDetails
9:00 AM – 11:00 AMSession 3 (S3) – Earth-Atmosphere SystemsGonzalo Mateo GarcíaAI for methane plume detection in satellite images
Jorge Pérez-AracilHeatwave reconstruction with a Deep Learning-based Analogue method
Breakout DiscussionsTopics entered and voted using Slido. Groups form based on the most-voted topics, moderated by the proposer. A note-taker is assigned for each group.
11:00 AM – 11:30 AMCoffee Break
11:30 AM – 1:00 PMSession 4 (S4) – Socio-Environmental SystemsIoannis AthanasiadisHybrid machine learning in agricultural modelling
Michele MeroniEO and machine learning in support to food security monitoring
1:00 PM – 1:30 PMSession 5 (S5) – VisionJorge Vila Tomás and Paula Daudén OliverModelling the Human Visual System with Parametric Neural Networks.
1:30 PM – 3:30 PMGroup Lunch
3:30 PM – 4:30 PMSession 5 (S5) – VisionDimosthenis KaratzasDocument Understanding in the era of Foundation Models
Nate MankovichA Flag Decomposition for Hierarchical Datasets
4:30 PM – 5:00 PMCoffee Break
5:00 PM – 6:00 PMBreakout DiscussionsTopics entered and voted using Slido.
Groups form based on the most-voted topics, moderated by the proposer.
A note-taker is assigned for each group.
9:00 PMDinner at Oslo Valencia RestaurantLocation: C/ Catalans, 8, 46001 València
Vegetarian and vegan cuisine in a historic 1850s building in El Carmen.
Focused on high-quality, seasonal, and locally sourced ingredients.

Day 3 – Wednesday, 15th January

TimeSessionPresenterDetails
9:00 AM – 11:00 AMSession 6 (S6) – Machine learningAnna Jungbluth3D Cloud Reconstruction via Geospatially-aware Masked Autoencoders
Pierre GentineReduced order models for prediction
Ismael Gomez TalalThe importance of explainable methods in artificial intelligence models, real use cases and ethical issues.
Breakout DiscussionsTopics entered and voted using Slido. Groups form based on the most-voted topics, moderated by the proposer. A note-taker is assigned for each group.
11:00 AM – 11:30 AMCoffee Break
11:30 AM – 1:00 PMSession 7 (S7) – Conclusions and Wrap-upPresentation TeamPresentation of breakout session outcomes.
Consolidated discussion on common themes of complexity and possible solutions across sessions.
Farewell and closing remarks.