Four threads, woven through two days.

Physics-informed climate emulators; ML-driven numerical weather prediction; deep learning for ocean biogeochemistry; AI-accelerated biogeochemical modeling; data-driven hydrology and ecosystem forecasting.

Foundation models for satellite and hyperspectral imagery; geospatial change detection and land-use classification; ML for planetary surface analysis and planetary exploration; deep learning for astronomy and astrophysics; AI/ML for solar physics, solar wind and magnetospheric forecasting; space weather nowcasting; AI/ML in satellite data processing.

Physics-informed neural networks and neural operators; AI/ML for theoretical and high-energy physics; the statistical and theoretical foundations of AI/ML; generalization and optimization theory of deep learning; information-theoretic frameworks for learning.

HPC and cloud computing for AI/ML; multi-step autonomous agent architectures; planning, memory, and tool-augmented reasoning; quantum machine learning; adversarial robustness and neural network verification; AI-driven threat detection and intrusion response; post-quantum cryptographic primitives for AI pipelines; privacy-preserving and quantum-resilient federated learning.