The course aims to acquire a deep understanding of models and technologies for knowledge representation and automated reasoning, with particular reference to knowledge graph models and their application in the financial and economic field. Specifically, the course will include an introductory part in which the fundamentals of data management, the logical perspective on data, uncertainty handling, main data models, and their query languages will be presented. The second part will focus on automated reasoning, delving into topics of knowledge modeling through ontologies, including elements related to the functioning of modern dedicated software systems. The entire course will be guided by real-life use cases applied in the financial and economic field, with an in-depth discussion of scenarios such as fraud detection, corporate governance, risk modeling, creditworthiness, and others.

1) In the first block, students are introduced to the fundamental concepts of AI, with the goal of presenting its different disciplines. A special focus will be on comparing inductive approaches, for instance, statistical learning, and deductive approaches, such as logic-based reasoning, in the perspective of their application in the world of KGs.2) In the second block, the course navigates into the data component of KGs, presenting methods and tools to represent, design, and manage them in the relational and graph-oriented models and languages, such as those specific for graphs. Moreover, the students will learn to manage data and intensionally specify knowledge in logic-based formalisms targeting reasoning systems. 3) The third block delves into the core of KGs, focusing on uncertainty in the data and, significantly, in the knowledge. Students will explore a spectrum of probabilistic reasoning approaches, including graph embeddings, graph neural networks, and Markov logic networks, which allow to capture uncertainty in complex relationships. 4) The fourth block enriches the course with real-world insights, honing in on financial knowledge graphs, for instance: modeling shareholding structures, exposure networks, transactional and payment data, real-world smart contracts, anti-money laundering settings, financial fraud detection, corporate governance, risk modeling, credit worthiness. Students are exposed to practical use cases, where reasoning is adopted to solve large-scale problems in these areas. The final grade will be a weighted average of the grades obtained with class participation and practical assignment (50%) and the final exam (50%). The final exam consists of a brief research proposal using methods covered in class to explore her own research field or extend a published research paper.

In presenza

Presentazione report scritto

P.D.2-2 - Giugno