Teacher: Paolo Torroni (DISI, University of Bologna) Teaching period: March/April 2027 Year of study: first or second Total hours: 15 Doctoral credits: 3 CD Final verification: Written exam with open questions
The course aims at providing a broad overview of LLMs from both a foundational and user perspective to PhD students from all areas of Computer Science and Engineering. It will be 80% theory and 20% practice, with a hands-on 3-hour session where students will be guided by prof. Ruggeri, through the implementation of a simple RAG/agentic pipeline using publicly available tools. The theory part will be developed within four 3-hour lessons, held by prof. Torroni, covering language representation, architectures, fine-tuning/alignment, responsible use, limitations, and benchmarking. The course material will include selected sections of the Jurafsky-Martin textbook and a selection of fundamental, relevant papers in computational linguistics and natural language processing.
Teacher: to be assigned Teaching period: May-June 2027 Year of study: first or second Total hours: 15 Doctoral credits: 3 CD Final verification: Report
The number of scientific articles published in Computer Science (and similar fields) increases steadily every year. This is mainly due to breakthroughs like Deep Learning, and, more recently, Large Language Models.
Paradoxically, researchers are struggling even more to reproduce published research. This issue affects all possible aspects of research, including methodology, data curation, approach comparison, and implementation.
In this course, we'll introduce and discuss the concept of 'reproducibility' in research. In particular, we'll overview current issues in research and existing attempts to address them. We'll focus on data curation, experimental setup, model comparison, and programming best practices.
Models and languages for service-oriented and cloud computing
Teacher: Mario Bravetti (DISI, University of Bologna) Teaching period: April-May 2027 Year of study: first or second Total hours: 15 Doctoral credits: 3 CD Final verification: Writing an in-depth report on a course topic
In the context of Service-Oriented and Cloud Computing one of the most prominent challenges is related with the composition of the software artifacts that need to correctly communicate and be properly configured. We review some of the models that have been proposed to perform a foundational study of the composition problem (choreographies/orchestrations/behavioural contracts/communicating automata), and we present languages and tools developed to support the realization of “correct-by-construction” compositions (implementing such models directly as programming languages or as types). Moreover, we address the problem of automatically deploying applications in the context of cloud computing, where it is necessary to provide the software artifacts with the required computing resources.
Teacher: to be assigned Teaching period: February-March or June-July 2027 Year of study: first or second Total hours: 15 Doctoral credits: 3 CD Final verification: Short essay on the usage of cluster computing for the specific PhD project, or oral examination.
This course introduces the fundamentals of High-Performance Computing (HPC) and cluster computing, explaining what they are, why they are used, and how they enable large-scale and high-speed computation. The course focuses on practical use of distributed cluster systems, including job scheduling, parallel execution, and resource management. It also introduces common libraries and frameworks for parallel and distributed computing (such as MPI and openMP), with an emphasis on how to execute applications on HPC systems in an efficient way.
Data Visualization for Researchers: theory and new approaches to enhance communication and dissemination
Teacher: Chiara Ceccarini (DISI, University of Bologna) Teaching period: June 2027 Year of study: first or second Total hours: 15 Doctoral credits: 3 CD Final verification: The assessment will be based on a project in which students apply data visualization principles and techniques to design and present an effective and ethically sound visualization for a real or realistic dataset.
This course aims to present the fundamentals of data visualization to empower students with the skills and insights necessary for creating effective visualizations that enhance communication and dissemination of data and scientific results. Focusing on the principles of data visualization, participants will explore diverse techniques to transform information into data representations, combining theoretical lessons with hands-on group activities. The theoretical sessions will cover fundamentals, techniques, best practices, and new trends in data visualization, while practical group activities will offer participants the chance to apply learned concepts in real-world scenarios. Finally, the course will present some ethical considerations about Data Visualization.
Teacher: Alessio Mora (DISI, University of Bologna) Teaching period: January-February 2027 Year of study: first or second Total hours: 20 Doctoral credits: 4 CD Final verification: The assessment will consist of either a practical project on an agreed topic or the preparation of a short research proposal (2 pages) on topics covered in the course.
Federated Learning (FL) is a decentralized paradigm that trains Machine/Deep Learning models without centralizing data. Privacy is preserved because data remains on user-controlled devices instead of being sent to a server. Edge devices (e.g., smartphones) contribute by training locally and periodically sending model updates to refine a global model. The course introduces widely used FL algorithms and key challenges, including heterogeneous data, communication/computation efficiency, privacy guarantees, and the right to be forgotten.
Teacher: Enrico Gallinucci (DISI, University of Bologna) Teaching period: February-March 2027 Year of study: first or second Total hours: 15 Doctoral credits: 3 CD Final verification: Students requiring an exam will be asked to produce an essay about possible intersections between their research interests and the topics/open issues presented in this course.
After introducing the motivations and principles behind NoSQL systems, the course presents the fundamental design principles and analyzes the main cost factors influencing data modeling decisions, including performance, storage occupation, and monetary costs in cloud environments. It reviews the state of the art in schema design for NoSQL databases and discusses in detail the innovative techniques and methodologies for identifying optimal schema designs under multiple cost factors. Finally, it addresses the challenges of refactoring databases in light of the evolution of the domain and query workload, evaluating the tradeoff between the opportunity of restoring efficiency and the effort of migrating the data.
Teacher: Nicolò Romandini (DISI, University of Bologna) Teaching period: October-November 2026 Year of study: first or second Total hours: 15 Doctoral credits: 3 CD Final verification: Oral presentation.
The course explores the transformation of digital identity from centralized, provider-controlled models to Self-Sovereign Identity (SSI), where individuals manage and control their own credentials. It introduces the core principles of digital identity and its evolution, providing a foundation for understanding challenges related to trust, privacy, and interoperability. SSI is a central focus, with a detailed examination of its architectures, standards, and enabling technologies, including decentralized identifiers (DIDs), verifiable credentials, cryptography, blockchain, and governance frameworks. Real-world applications, including European Union initiatives, illustrate the practical adoption and impact of SSI in building user-centric, privacy-preserving digital identity ecosystems.