Sessions
54th International Liège Colloquium on Ocean Dynamics | 8 to 12 May 2023
1. Learning from Numerical Models
1.a Focus on model development
1.b Pre- and post-processing model output
- Surrogate models and forecasting using ML
- Model calibration and tuning
- New data-driven parameterization in models
- Forecast of extreme values
- Improving machine learning or traditional techniques by incorporating dynamics knowledge
2. Learning from Observations
- Merge different datasets, data fusion, and multivariate analysis
- Retrieval algorithm of satellite data
- Machine learning-based data QA/QC
- Reconstruction of missing data and inference of subsurface information from satellite and in situ data
- Dimensionality issues
3. Cross-cutting approaches and integration
- Enhancing the spatial or temporal resolution, or the accuracy of datasets
- Probabilistic approaches in machine learning, including uncertainty quantification
- Quantify and improve the accuracy of data products including bias estimation
- Extracting information, patterns and features from ocean data (observations and models) and feature identification
- Adaptive sampling
- Data analytics for systematic analysis of ocean data
- Interpretable and Explainable AI techniques
- Physics-informed machine learning
- Data assimilation with machine learning and optimal ensemble predictions
- Integration of statistical models
updated on 10/2/23
