Mattia Pujatti
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Mattia Pujatti has a total of 12 co-authors. Here are the top co-authors by number of shared publications:
- Marco Cristoforetti
INFN-TIFPA
3 shared publications - Andrea Di Luca
INFN-TIFPA
2 shared publications - Davide Calzà
Fondazione Bruno Kessler, Trento, Italy
2 shared publications Andrea Gobbi
Fondazione Bruno Kessler, Trento, Italy
1 shared publication- Andrea Gobbi
Fondazione Bruno Kessler
1 shared publication - Federico Monegaglia
Università degli Studi di Trento
1 shared publication
Latest Publications
- 24 October 2025
- Published by Elsevier in Expert Systems with Applications
- Vol. 299
Deep learning-based control optimization for glass bottle forming
- 24 October 2025
- Published by Elsevier in Expert Systems with Applications
- Vol. 299
Deep learning-based control optimization for glass bottle forming
- Open Access
Deep Learning-Based Control Optimization for Glass Bottle Forming
In glass bottle manufacturing, precise control of forming machines is critical for ensuring quality and minimizing defects. This study presents a deep learning-based control algorithm designed to optimize the forming process in real production environments. Using real operational data from active manufacturing plants, our neural network predicts the effects of parameter changes based on the current production setup. Through a specifically designed inversion mechanism, the algorithm identifies the optimal machine settings required to achieve the desired glass gob characteristics. Experimental results on historical datasets from multiple production lines show that the proposed method yields promising outcomes, suggesting potential for enhanced process stability, reduced waste, and improved product consistency. These results highlight the potential of deep learning to process control in glass manufacturing.
Open AccessDeep Learning-Based Control Optimization for Glass Bottle Forming
In glass bottle manufacturing, precise control of forming machines is critical for ensuring quality and minimizing defects. This study presents a deep learning-based control algorithm designed to optimize the forming process in real production environments. Using real operational data from active manufacturing plants, our neural network predicts the effects of parameter changes based on the current production setup. Through a specifically designed inversion mechanism, the algorithm identifies the optimal machine settings required to achieve the desired glass gob characteristics. Experimental results on historical datasets from multiple production lines show that the proposed method yields promising outcomes, suggesting potential for enhanced process stability, reduced waste, and improved product consistency. These results highlight the potential of deep learning to process control in glass manufacturing.
- Open Access
- 4 April 2025
- Published by Springer Nature in Journal of Big Data
- Vol. 12 (1) , 1-33
Mopidip: a modular real-time pipeline for machinery diagnosis and prognosis based on deep learning algorithms
- 4 April 2025
- Published by Springer Nature in Journal of Big Data
- Vol. 12 (1) , 1-33
Open AccessMopidip: a modular real-time pipeline for machinery diagnosis and prognosis based on deep learning algorithms
- Open Access
- 19 July 2024
- Published by MDPI AG in Preprints.org
MoPiDiP: A Modular Real-Time Pipeline for Machinery Diagnosis and Prognosis Based on Deep Learning Algorithms
Condition monitoring is a crucial process for ensuring industrial assets' reliability and operational efficiency. In the age of the digital industry, AI-based data-driven condition monitoring is proving extremely effective in detecting potential issues before they escalate into major problems, thereby reducing downtime, minimizing maintenance costs, and extending the lifespan of the equipment. The availability of tools that can enable the operationalization of these data-driven solutions is, therefore, critical. In this direction, this work proposes a comprehensive, modular, and scalable pipeline covering all the steps from the data acquisition to the AI model training and inference phases. The tool integrates the data acquisition and processing steps with a configurable feature extraction phase. Moreover, the system also integrates deep learning algorithms for diagnosis and prognosis, including a domain adaptation stage to permit transfer learning and increase generalizability. In addition, it features a communication system using MQTT, which allows for an online data stream to enable real-time monitoring and maintenance. The overall infrastructure was deployed in actual industrial settings and tested in a real-time experiment, demonstrating the proposed approach's validity.
- 19 July 2024
- Published by MDPI AG in Preprints.org
Open AccessMoPiDiP: A Modular Real-Time Pipeline for Machinery Diagnosis and Prognosis Based on Deep Learning Algorithms
Condition monitoring is a crucial process for ensuring industrial assets' reliability and operational efficiency. In the age of the digital industry, AI-based data-driven condition monitoring is proving extremely effective in detecting potential issues before they escalate into major problems, thereby reducing downtime, minimizing maintenance costs, and extending the lifespan of the equipment. The availability of tools that can enable the operationalization of these data-driven solutions is, therefore, critical. In this direction, this work proposes a comprehensive, modular, and scalable pipeline covering all the steps from the data acquisition to the AI model training and inference phases. The tool integrates the data acquisition and processing steps with a configurable feature extraction phase. Moreover, the system also integrates deep learning algorithms for diagnosis and prognosis, including a domain adaptation stage to permit transfer learning and increase generalizability. In addition, it features a communication system using MQTT, which allows for an online data stream to enable real-time monitoring and maintenance. The overall infrastructure was deployed in actual industrial settings and tested in a real-time experiment, demonstrating the proposed approach's validity.