Machine Learning On Graphs

1. Knowledge and understanding
Acquire a solid theoretical background in machine learning techniques for graph-structured data, including Graph Neural Networks, embedding algorithms, and models for heterogeneous and temporal graphs.
2. Applying knowledge and understanding
Develop practical skills in implementing, using, and evaluating graph learning algorithms with modern libraries (NetworkX, DGL, PyTorch Geometric), and applying them to real-world problems in various domains.
3. Making judgements
Gain the ability to critically assess, select, and adapt appropriate graph learning methods based on application context, graph structure, and analysis goals.
4. Communication skills
Effectively present design choices, methodologies, and results, both orally and in writing, including through the discussion of relevant scientific papers.
5. Learning skills
Develop autonomy in learning advanced methods and staying up to date through scientific literature, fostering the ability to acquire new models and technologies in the field.

Describe and explain the main graph-based machine learning algorithms, including approaches without feature learning and graph representation learning models.
Implement and apply graph embedding techniques using specialized Python libraries.
Evaluate algorithm performance on real datasets, selecting suitable metrics and optimizing models accordingly.
Design solutions for practical problems involving graph-structured data, including heterogeneous and temporal networks.
Critically analyze scientific papers and implement the described methods, discussing their advantages, limitations, and applicability.
Clearly communicate methodological choices and results, both in written reports and oral presentations.