Responsible AI Engineering
Lecturers: Humberto Torres Marques-Neto (PUC Minas, Brasil)
Contact: humberto@pucminas.br
Schedule: December 15 & 16 & 17 & 18, 2026, from 11am to 1pm and from 2pm to 4 pm – sala seminari est
The emergence of AI software engineering marks a transformation in traditional software development practices. Unlike conventional approaches – centered on well-defined requirements and deterministic algorithms – AI development embraces iterative processes such as model training, validation, and deployment. These stages are inherently probabilistic and data-driven, emphasizing adaptation and continuous learning from historical datasets and user interactions. Moreover, the widespread adoption of AI technologies in recent years, together with the integration of AI-assisted tools and Large Language Models (LLMs), has introduced new paradigms in how software systems are designed, developed, and maintained. Effectively incorporating AI techniques and tools within the software lifecycle requires engineers to embrace agile principles, prioritize data quality and governance, and account for ethical dimensions to promote fairness, reliability, privacy, transparency, sustainability, accountability, and explainability. This ongoing evolution from software engineering to Responsible AI engineering represents a pivotal step toward leveraging AI’s full potential while fostering responsible and sustainable innovation. According to the Software Engineering Institute of the Carnegie Mellon University – SEI@CMU, the AI Engineering proposes the combining of system engineering, software engineering, computer science, and human-centered design to create AI Systems to be scalable, robust, and secure for running in complex contexts and with a minimum predictability of maintenance and budget. Moreover, in addition to the importance of building high-quality with the high-productivity AI-based software systems, software teams should be aware of and consider best practices for the responsible use of AI in different domains. This course will present concepts, discuss case studies, and debate research opportunities about Responsible AI Engineering. In the course, students will be invited to analyze and discuss issues regarding the software processes and their ethical implications, AI-based software requirement engineering using LLMs, defining software architecture for AI systems, software quality, deployment, as well as the usage of AI to code and test AI Systems. We expected students to apply the discussed concepts and present a final seminar using data from large-scale systems, such as online social network data and open government data, within the broad context of an AI System.
A basic and concise introduction to Topological Data Analysis
Lecturer: Patrizio Frosini (UNIPI)
Contact: patrizio.frosini@unipi.it
Schedule: January 11 & 13 & 15 & 18 & 20 & 22 & 25 & 26, 2027, from 11am to 1pm – sala seminari est
Topological Data Analysis (TDA) is a mathematical framework focused on studying and quantifying the “shape” of data. Its primary goal is to describe and measure the similarity in datasets by using distances, particularly when equivalences are defined through geometric transformations. Additionally, TDA is highly effective for reducing the dimensionality of data, making it easier to analyze and compare. It can also be utilized in geometric machine learning, and its approach can be applied to a wide range of data types, including time series, 2D and 3D objects, and point clouds. Throughout the course, fundamental concepts required for a basic understanding of TDA will be introduced, with a focus on practical and computational examples, rather than formal mathematical theory.
Topics:
- Equivalence and non-equivalence of data with respect to the action of a group of transformations.
- Simplicial complexes as a generalization of the concept of a graph and as a geometric representation of data described by point clouds in Euclidean spaces.
- Simplicial homology groups as a method for representing the “shape” of a simplicial complex derived from a point cloud.
- The need to adapt homology to the observer’s point of view and the presence of noise: an introduction to persistent homology and persistence diagrams.
- Stability of persistence diagrams in the presence of noise.
- Applications of persistent homology.
- From the shape of data to the shape of observers: the concept of a Group Equivariant Non-Expansive Operator (GENEO).
- The problem of approximating observers in the space of GENEOs.
- GENEOs as a geometric method for reducing the number of parameters in neural networks and increasing their interpretability.
- Applications of GENEO Theory to Geometric Machine Learning.
Distributed Ledger Technology: data management and analysis
Lecturers: Damiano Di Francesco Maesa (UNIPI), Matteo Loporchio (UNIPI)
Contact: damiano.difrancesco@unipi.it
Schedule: January 12 & 14 & 19 & 21, 2027, from 11am to 1pm and from 2pm to 4pm – sala seminari est
The goal of this course is to present how data are managed (represented, secured, and retrieved) in Distributed Ledger Technology (DLT) based systems, and how data can be analysed to study the ecosystems they support. The course will start with an introduction to the concepts behind DLT, including its main innovative applications. One lecture will be dedicated to the critical analysis of an AI-based application of DLT, highlighting the novel capabilities offered by this technology. We will then present the Ethereum and Bitcoin protocols, outlining how they manage their internal transaction data. The same data will be the focus of the following lectures showcasing how to represent and analyse them through graphs. The course will close by presenting how authenticated data structures can be leveraged to enhance such data management.
Computational Modeling for Systems Biology
Lecturers: Paolo Milazzo (UNIPI), Silvia Galfré (UNIPI)
Contact: paolo.milazzo@unipi.it
Schedule: February 2 & 3 & 4 & 9 & 10 & 11, 2027, from 10am to 1pm – sala seminari est
The course will deal with several aspects of the in-silico analysis of dynamical properties of biological systems. We will focus, in particular, on mechanistic modeling approaches aiming at creating executable representations of the biological mechanisms and processes underlying cell functioning. After providing a few notions of biochemistry and cell biology, we will examine modeling methods for gene regulatory networks with particular emphasis on Boolean network models and rule-based approaches. Next, we will present approaches suitable for the analysis of metabolic and cell-signaling processes, ranging from differential equations, to stochastic modeling and simulation methods, to hybrid approaches. Finally, we will briefly survey emerging methods in computational structural biology, such as methods for protein structure prediction and molecular dynamics simulation, and we discuss how these techniques could be integrated with the previous ones in order to evaluate the impact of protein mutations on cell functioning.
Pathways to Green ICT
Lecturers: Antonio Brogi (UNIPI), Stefano Forti (UNIPI)
Contact: antonio.brogi@unipi.it
Schedule: February 2027
The course aims at introducing students to the fundamentals of Green ICT, providing them with a toolbox to consider sustainability aspects in their research. The course will introduce:
- The concepts of sustainability and the types of environmental impact of the lifecycle of ICT systems (power consumption, carbon emissions, e-waste)
- Methodologies to assess the environmental impact of ICT systems (from production to operation and maintenance to disposal)
- Methodologies to reduce the environmental impact of ICT systems (orthogonality of QoS and environmental goals, hardware selection and PUE reduction, energy-aware programming, green software engineering, energy-aware system deployment)
- Use cases and open research challenges
Governing AI Sprawl: Preventing Performance Degradation, Quality Erosion and Security Failures in AI-Gen Systems
Lecturers: Antonello Calabrò (CNR-ISTI)
Contact: antonello.calabro@isti.cnr.it
Period: March-April 2027
The increasing ability of generative AI systems and autonomous agents to produce source code, tests, configurations, documentation, schemas, pipelines, infrastructure definitions, and complete software components is dramatically reducing the cost of software generation.
However, faster generation does not necessarily produce faster, better, or safer software systems. When the production of new artifacts exceeds an organisation’s capacity to discover, understand, review, integrate, secure, maintain, consolidate, and retire them, local productivity gains can generate system-wide degradation.
The resulting phenomenon is AI sprawl, which may contribute to declining software delivery performance, slower maintenance and change propagation, architectural erosion, uncontrolled dependencies, and progressive quality degradation.
It may also lead to duplicated or conflicting functionality, reduced reliability and maintainability, weakened security assurance, an expanded attack surface, insecure dependencies, and increased software supply-chain risk.
The course introduces AI sprawl as a software engineering and governance problem and provides students with the theoretical, methodological, and technical instruments required to recognise and reduce its different manifestation defining governance models for joint human–agent software development.
Contents:
- Why AI Sprawl Emerges: From Local Productivity to Global Performance Degradation
- Formal Foundations of AI Sprawl
- Architectural and Functional Sprawl: Quality Erosion Through Duplication and Divergence
- Artifact, Data, and Process Sprawl: Opacity, Traceability Loss, and Reliability Degradation
- Tool, Agent, Security, and Organisational Sprawl
- Detection, Metrics, Risk Assessment, and Open-Source Tooling
- Prevention and Remediation: Restoring Performance, Quality, and Control
Programming Tools and Techniques in the Pervasive Parallelism Era
Lecturers: Marco Danelutto (UNIPI), Patrizio Dazzi (UNIPI)
Contact: marco.danelutto@unipi.it
Schedule: May 5 & 6 & 7 & 10 & 11 & 12 & 13 & 14, 2027, from 11am to 1pm – sala seminari est
The course covers techniques and tools (already existing or that are in the process of being moved to mainstream) suitable to support the implementation of efficient parallel/distributed applications targeting small scale parallel systems as well as larger scale parallel and distributed systems, possibly equipped with different kind of accelerators. The course follows a methodological approach to provide a homogeneous overview of classical tools and techniques as well as of new tools and techniques specifically developed for new, emerging architectures and applicative domains. Perspectives in the direction of reconfigurable coprocessors and domain-specific architectures will also be covered.
3D Geometry Representation and Processing for Deep Learning
Lecturers: Paolo Cignoni, Massimiliano Corsini, Daniela Giorgi, Luigi Malomo (CNR-ISTI)
Contact: daniela.giorgi@isti.cnr.it
Schedule May-June 2027
Computer Graphics and Geometry Processing are the main disciplines dealing with 3D data such as meshes and point clouds. In turn, Artificial Intelligence and Deep Learning are fundamental paradigms to manage visual data. Nevertheless, applying traditional learning paradigms on 3D data requires rethinking architectural building blocks designed for 2D images, such as convolution and pooling operators, as well as attention layers.
In this course, we will introduce different representations for 3D data, and basic geometry processing techniques that intervene in deep learning pipelines (sampling, remeshing, conversion, …). Then, we will introduce methods able to learn tasks on 3D data. We will describe different architectures to process complex geometric domains, and the novel mechanisms introduced in the literature to preserve by design their intrinsic properties. Examples include graph learning techniques, augmented with geometric and topological information; attention modules to process unordered point sets and mesh data; transformer-like architectures for unstructured data.
In the second part of the course, we will discuss different applications where the interplay between Computer Graphics/Geometry Processing and Deep Learning is opening up to exciting results, including Computational Fabrication, Architectural Geometry, and Environmental Monitoring.
From Search to Generation: Modern Information Access
Lecturers: Franco Maria Nardini, Cosimo Rulli, Salvatore Trani (CNR-ISTI), Rossano Venturini (UNIPI)
Contact: francomaria.nardini@isti.cnr.it
Schedule: June 22 & 23 & 24 & 25, 2027, from 9am to 1pm – sala seminari est
This PhD course examines the state of the art and open challenges in modern information access, spanning three interconnected areas: i) indexing, ii) query processing and ranking, and iii) generation after retrieval. The course introduces foundational and advanced techniques for indexing and querying large-scale Web collections, with particular emphasis on query processing and the role of machine learning in improving ranking effectiveness. It covers supervised learning-to-rank methods and analyzes the trade-offs among effectiveness, latency, computational cost, and memory consumption.Building on these foundations, the course explores how transformer-based large language models are reshaping information access through neural retrieval, reranking, and retrieval-augmented generation. It examines key challenges arising when retrieved information is used to generate answers, including grounding, factuality, source attribution, hallucination, freshness, bias, evaluation, and end-to-end efficiency. The course also discusses the evolving relationship between traditional ranked-result search and generative answer systems, highlighting open research questions across retrieval and generation. Hands-on sessions will give participants practical experience with indexing, retrieval, ranking, and retrieval-augmented generation using public Web collections.