Doctoral Candidate

Donato Cerciello, MSc

Donato_Cerciello received his Master’s degree in Mathematical Engineering from University ofNaples Federico II. His academic background includes and implementing the design of a generative framework of digital twins, based on algorithms for Time Series Generation, using Convolutional Neural Networks enhanced with Long-Short Term Memory, to enable simulation and forecasting with accuracy for virtual testing of mobility strategies before real-world implementation, closely aligning with the goals of the TUAI project.

Donato_Cerciello_Self-presentaion.pdf 

Main Supervisors: David Camacho (UPM), Javier Huertas Tato (UPM) 

Co-Supervisors: Francesco Piccialli (UNINA), Dariusz Mrozek (SUT), Jia-Chun Lin (NTNU) 

R&D cooperation: GMV

Objectives: advancing deep learning methodologies for unsupervised pattern discovery in temporal data. 

The main goal is to design and evaluate deep clustering models capable of capturing complex temporal dynamics while providing transparent and interpretable explanations of their results.

The project will address major challenges in time series clustering, including robustness to non-stationarity, domain shifts, and incomplete data. A strong emphasis will be placed on explainability, leveraging visualization tools and language-based interpretability to make the outcomes of clustering models more understandable to both researchers and end-users. The research will also explore hybrid architectures combining deep representation learning with visual analytics to interpret temporal evolution and detect anomalies.

Expected Results: 

A robust prototype of explainable deep clustering for time series validated on real-world data.

A framework for visual and quantitative evaluation of temporal cluster quality and robustness.

Publications in top-tier AI venues (e.g.ICML, ECML PKDD, Information Fusion).

Contributions to the TUAI objectives in WP2, advancing explainable and trustworthy AI for time series analysis.

Applied research: The primary outcome will be the innovation of XAI techniques and deep learning models, to effectively identify anomalies and patterns within time series data, which will be applied to diverse industrial domains, such as space and industry including applied research with GMV , and will serve to showcase the developed models and techniques.

Planned secondments: UNINA (4 months); SUT(4 months); NTNU(4 months)                    

Enrolment in Doctoral degree: UPM