Barbara Toniella Corradini
Postdoctoral Researcher
Istituto Italiano di Tecnologia (IIT), Genoa, Italy
📧 barbaracorradini5@gmail.com
🔗 GitHub
🎓 Google Scholar
🧬 ORCID
Research Statement
I am a Postdoctoral Researcher at the Istituto Italiano di Tecnologia (IIT) working on Artificial Intelligence and Deep Learning, with a particular emphasis on generative models, computer vision, and multimodal learning.
My research focuses on understanding and exploiting the internal structure, semantic properties,
and training dynamics of large generative architectures, including GANs, diffusion models,
and vision–language models.
I combine theoretical analysis with applied machine learning, targeting scientific,
biomedical, and real-world applications.
A key aspect of my work is the study of model interpretability and evaluation, using tools such as topological data analysis and statistical methods to analyze representation learning and training behavior.
Current Position
Postdoctoral Researcher
Istituto Italiano di Tecnologia (IIT), Genoa
2025 – Present
Previous Positions
Research Fellow
University of Siena · 2025
Visiting Researcher
Sorbonne Université (ISIR Lab), Paris · 2023
Education
PhD in Smart Computing (Doctor Europaeus)
University of Florence & University of Siena
2021 – 2024
MSc in Computer and Automation Engineering
University of Siena · 110/110 cum laude
BSc in Computer and Information Engineering
University of Siena
Publications
Selected Publications
-
Training Dynamics of GANs Through the Lens of Persistent Homology
Neurocomputing, 2025 -
FreeSeg-Diff: Training-Free Open-Vocabulary Segmentation with Diffusion Models
IJCNN 2025 -
DiffProps: Is Semantics Preserved within a Diffusion Model?
KES 2024
Teaching & Academic Service
- Session Chair — IJCNN 2025
- Reviewer for IJCNN, ESANN, ECML PKDD, ICANN, Neurocomputing, Soft Computing, and more
- Courses in Programming, Databases, Artificial Intelligence (Italy & International)