Machine Learning for Environmental Sciences Course Overview
I gained hands-on experience and theoretical knowledge across a broad range of topics regarding machine learning for environmental sciences. Below is a detailed week-by-week summary of the course content and the practical exercises I completed.
Week 1: Introduction to Machine Learning
- Lecture: Definition and history of machine learning.
- Exercise: Python programming basics and analyzing global warming data (Exercise01.ipynb).
Week 2: Foundations of Machine Learning
- Lecture: Core principles and foundational concepts.
- Exercise: Continued work on Python and global warming data analysis (Exercise01.ipynb).
Week 3: Regularization and Hyperparameters
- Lecture: Techniques such as Ridge and LASSO for regularization, and understanding hyperparameters.
- Exercise: Exploring the curse of dimensionality (Exercise02.ipynb).
Week 4: Ensemble Methods
- Lecture: Ensemble regression and classification methods, including Random Forests.
- Exercise: Implementing random forest regression (Exercise03.ipynb).
Week 5: Explainable AI
- Lecture: Introduction to SHAP values and feature importances for explainable AI.
- Exercise: Explaining data-driven models (Exercise04.ipynb).
Week 6: Deep Learning
- Lecture: Fundamentals of deep learning and convolutional neural networks (CNNs).
- Exercise: Tuning neural networks for improved performance (Exercise05.ipynb).
Week 7: Recurrent Neural Networks (RNNs)
- Lecture: Understanding RNNs and Long Short-Term Memory (LSTM) networks.
- Exercise: Time series forecasting using RNNs (Exercise06.ipynb).
Week 8: Gaussian Process Regression
- Lecture: Techniques for Gaussian process regression.
- Exercise: Climate model emulation (Exercise07.ipynb).
Week 9: Unsupervised Learning (I)
- Lecture: Principal Component Analysis (PCA) and k-means clustering.
- Exercise: Dimension reduction techniques (Exercise08.ipynb).
Week 10: Unsupervised Learning (II)
- Lecture: Self-organizing maps and Varimax rotation.
- Exercise: Analyzing weather maps (Exercise09.ipynb).
Week 11: Causal Discovery
- Lecture: Techniques for discovering causal relationships.
- Exercise: From correlation to causation (Exercise10.ipynb).
Week 12: Transformer Architectures
- Lecture: Introduction to transformer architectures and attention mechanisms.
- Exercise: Exploring transformer models (Exercise11.ipynb).
Week 13: Anomaly Detection
- Lecture: Methods for detecting anomalies in data.
- Recap: Reviewing key concepts and practical applications.
Week 14: Self-Supervised Learning (Optional)
- Lecture: Introduction to self-supervised learning techniques.
- Recap: Final review and consolidation of learning.