From financial fraud detection to drug discovery, from online recommendation systems to cybersecurity, many of the most impactful problems today are fundamentally network problems.

Modern systems are not collections of isolated data points—they are webs of interactions. Transactions form financial graphs. Proteins interact in biological networks. Information spreads through social systems. Infrastructure, supply chains, and communication platforms operate as interconnected structures. In such environments, relationships are often more informative than individual entities.

Graphs provide the mathematical and computational framework for understanding these systems.

Graph Mining and Applications studies how to analyze, model, and learn from relational data at scale. This new course bridges theoretical computer science and modern machine learning, combining algorithmic rigor with practical relevance. It develops the tools needed to reason about connectivity, influence, clustering, robustness, and structural patterns—concepts that directly underpin applications in finance, biology, social systems, knowledge representation, and beyond.

Unlike areas such as language and vision, graph-based modeling does not yet have a single unifying foundation model. The structural diversity of networks across domains makes this both challenging and intellectually exciting. Advances in this space have immediate implications for real-world systems that shape economies, health, and information flow.

This course prepares students to think structurally: to understand how networks behave, how to design algorithms for them, and how to leverage modern learning methods responsibly and effectively.

Announcements

Remember to register your email; details will be given in the first week of classes.

Classes start on Monday, February 23.

Topics that we will cover

• Properties of real-world networks
• Generative models for complex networks
• Community detection and spectral techniques
• Cascading processes, epidemics, and influence maximization
• Influence, homophily, and opinion dynamics
• Graph neural networks and graph transformers
• Graphlets and network motifs
• Temporal graphs and temporal motifs
• Hypergraphs and higher-order interactions
• Knowledge graphs
• Graph summarization, sampling, and generation

We will give particular emphasis on applications such as biological, chemical, social, financial networks and how the can be used to solve real-world problems.

Instructor

Aris Anagnostopoulos, Sapienza University of Rome

When and where:

Monday 13.00–15.00, Via Ariosto 25, Room A3

Thursday 13.00–16.00, Via Ariosto 25, Room A5

Office hours

You can use the office hours for any question regarding the class material, general questions on networks, the meaning of life, pretty much anything. Send an email to the instructors for arrangement.

Textbook and references

This is an advanced level course for which there does not exist a textbook containing all the material. We will make available notes, slides, book chapters, papers, and so on as we proceed.

Evaluation format

The recommended way to be evaluated is to do a project or a study in a given topic, which will be presented in class. Detail will be given during the first lectures.

Alternatively, there is also the option of a written exam for both parts.

Syllabus