Andrea Rosani
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Andrea Rosani has a total of 34 co-authors. Here are the top co-authors by number of shared publications:
- Giuseppe Di Fatta
Free University of Bozen Bolzano Italy
4 shared publications - Thomas Borsani
Free University of Bozen-Bolzano, Bozen-Bolzano, Italy
4 shared publications - Andrea Massa
DICAM - Department of Civil, Environmental, and Mechanical Engineering, ELEDIA Research Center (ELEDIA@UniTN - University of Trento), Trento, Italy
2 shared publications D. Franceschini
Department of Information and Communication Technology, University of Trento, Trento, Italy
2 shared publications- Giuseppe Nicosia
University of Catania, Catania, Italy
2 shared publications M. Pastorino
Department of Biophysical and Electronic Engineering, University of Genoa, Genoa, Italy
2 shared publications
Latest Publications
- Open Access
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 38-41
High-quality data augmentation for code comment classification
Code comments serve a crucial role in software development for documenting functionality, clarifying design choices, and assisting with issue tracking. They capture developers’ insights about the surrounding source code, serving as an essential resource for both human comprehension and automated analysis. Nevertheless, since comments are in natural language, they present challenges for machine-based code understanding. To address this, recent studies have applied natural language processing (NLP) and deep learning techniques to classify comments according to developers’ intentions. However, existing datasets for this task suffer from size limitations and class imbalance, as they rely on manual annotations and may not accurately represent the distribution of comments in real-world codebases. To overcome this issue, we introduce new synthetic oversampling and augmentation techniques based on high-quality data generation to enhance the NLBSE’26 challenge datasets. Our Synthetic Quality Oversampling Technique and Augmentation Technique (Q-SYNTH) yield promising results, improving the base classifier by \(2.56\%\) . The source code is publicly available at https://github.com/ThomBors/NLBSE2026.- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 38-41
Open AccessHigh-quality data augmentation for code comment classification
Code comments serve a crucial role in software development for documenting functionality, clarifying design choices, and assisting with issue tracking. They capture developers’ insights about the surrounding source code, serving as an essential resource for both human comprehension and automated analysis. Nevertheless, since comments are in natural language, they present challenges for machine-based code understanding. To address this, recent studies have applied natural language processing (NLP) and deep learning techniques to classify comments according to developers’ intentions. However, existing datasets for this task suffer from size limitations and class imbalance, as they rely on manual annotations and may not accurately represent the distribution of comments in real-world codebases. To overcome this issue, we introduce new synthetic oversampling and augmentation techniques based on high-quality data generation to enhance the NLBSE’26 challenge datasets. Our Synthetic Quality Oversampling Technique and Augmentation Technique (Q-SYNTH) yield promising results, improving the base classifier by \(2.56\%\) . The source code is publicly available at https://github.com/ThomBors/NLBSE2026. - Open Access
High-quality data augmentation for code comment classification
Code comments serve a crucial role in software development for documenting functionality, clarifying design choices, and assisting with issue tracking. They capture developers' insights about the surrounding source code, serving as an essential resource for both human comprehension and automated analysis. Nevertheless, since comments are in natural language, they present challenges for machine-based code understanding. To address this, recent studies have applied natural language processing (NLP) and deep learning techniques to classify comments according to developers' intentions. However, existing datasets for this task suffer from size limitations and class imbalance, as they rely on manual annotations and may not accurately represent the distribution of comments in real-world codebases. To overcome this issue, we introduce new synthetic oversampling and augmentation techniques based on high-quality data generation to enhance the NLBSE'26 challenge datasets. Our Synthetic Quality Oversampling Technique and Augmentation Technique (Q-SYNTH) yield promising results, improving the base classifier by $2.56\%$.Open AccessHigh-quality data augmentation for code comment classification
Code comments serve a crucial role in software development for documenting functionality, clarifying design choices, and assisting with issue tracking. They capture developers' insights about the surrounding source code, serving as an essential resource for both human comprehension and automated analysis. Nevertheless, since comments are in natural language, they present challenges for machine-based code understanding. To address this, recent studies have applied natural language processing (NLP) and deep learning techniques to classify comments according to developers' intentions. However, existing datasets for this task suffer from size limitations and class imbalance, as they rely on manual annotations and may not accurately represent the distribution of comments in real-world codebases. To overcome this issue, we introduce new synthetic oversampling and augmentation techniques based on high-quality data generation to enhance the NLBSE'26 challenge datasets. Our Synthetic Quality Oversampling Technique and Augmentation Technique (Q-SYNTH) yield promising results, improving the base classifier by $2.56\%$. - Open Access
- 19 November 2025
- Published by CERN in Open Research Europe
- Vol. 5, 352
AI4SWEng manifest: Empowering the future of software engineering with advanced trustworthy artificial intelligence
AI4SWEng, a Horizon Europe project that will be active from 2025 until 2028, unites 15 leading partners across the European Union, Switzerland and Turkey, combining experts in Model-Driven Software Engineering and trustworthy AI, with a special focus on applying Large Language Models. The project addresses complex challenges in industries such as healthcare, cyber-physical systems, and electric vehicles, focusing on multi-architectural and resource-constrained systems. Our mission is to transform agile software development by leveraging AI to boost efficiency, reliability, and security while ensuring ethical and regulatory compliance. The goal is to deliver scalable, sustainable, and socially responsible solutions that accelerate time-to-market without compromising quality. The AI4SWEng project will deliver an AI-powered software engineering suite providing end-to-end support for the software lifecycle, from code generation and advanced debugging to security, energy efficiency, and project management. By reducing pain points and enhancing productivity, the suite/platform aims to reduce developer stress, foster creativity, and improve job satisfaction. With a commitment to user-centred design and advanced prompt engineering, we empower developers to harness the full potential of AI. In this paper the AI4SWEng project consortium presents a novel and strategic approach proposed by the project consortium that will shape the future of software engineering, driving innovation and paving the way for more agile, intelligent, and sustainable software development.
- 19 November 2025
- Published by CERN in Open Research Europe
- Vol. 5, 352
Open AccessAI4SWEng manifest: Empowering the future of software engineering with advanced trustworthy artificial intelligence
AI4SWEng, a Horizon Europe project that will be active from 2025 until 2028, unites 15 leading partners across the European Union, Switzerland and Turkey, combining experts in Model-Driven Software Engineering and trustworthy AI, with a special focus on applying Large Language Models. The project addresses complex challenges in industries such as healthcare, cyber-physical systems, and electric vehicles, focusing on multi-architectural and resource-constrained systems. Our mission is to transform agile software development by leveraging AI to boost efficiency, reliability, and security while ensuring ethical and regulatory compliance. The goal is to deliver scalable, sustainable, and socially responsible solutions that accelerate time-to-market without compromising quality. The AI4SWEng project will deliver an AI-powered software engineering suite providing end-to-end support for the software lifecycle, from code generation and advanced debugging to security, energy efficiency, and project management. By reducing pain points and enhancing productivity, the suite/platform aims to reduce developer stress, foster creativity, and improve job satisfaction. With a commitment to user-centred design and advanced prompt engineering, we empower developers to harness the full potential of AI. In this paper the AI4SWEng project consortium presents a novel and strategic approach proposed by the project consortium that will shape the future of software engineering, driving innovation and paving the way for more agile, intelligent, and sustainable software development.
- 4 October 2025
- Published by Springer Nature
- p. 95-111
Gradient Similarity Surgery in Multi-task Deep Learning
- 4 October 2025
- Published by Springer Nature
- p. 95-111
Gradient Similarity Surgery in Multi-task Deep Learning
- Open Access
Gradient Similarity Surgery in Multi-Task Deep Learning
The multi-task learning ($MTL$) paradigm aims to simultaneously learn multiple tasks within a single model capturing higher-level, more general hidden patterns that are shared by the tasks. In deep learning, a significant challenge in the backpropagation training process is the design of advanced optimisers to improve the convergence speed and stability of the gradient descent learning rule. In particular, in multi-task deep learning ($MTDL$) the multitude of tasks may generate potentially conflicting gradients that would hinder the concurrent convergence of the diverse loss functions. This challenge arises when the gradients of the task objectives have either different magnitudes or opposite directions, causing one or a few to dominate or to interfere with each other, thus degrading the training process. Gradient surgery methods address the problem explicitly dealing with conflicting gradients by adjusting the overall gradient trajectory. This work introduces a novel gradient surgery method, the Similarity-Aware Momentum Gradient Surgery (SAM-GS), which provides an effective and scalable approach based on a gradient magnitude similarity measure to guide the optimisation process. The SAM-GS surgery adopts gradient equalisation and modulation of the first-order momentum. A series of experimental tests have shown the effectiveness of SAM-GS on synthetic problems and $MTL$ benchmarks. Gradient magnitude similarity plays a crucial role in regularising gradient aggregation in $MTDL$ for the optimisation of the learning process.Open AccessGradient Similarity Surgery in Multi-Task Deep Learning
The multi-task learning ($MTL$) paradigm aims to simultaneously learn multiple tasks within a single model capturing higher-level, more general hidden patterns that are shared by the tasks. In deep learning, a significant challenge in the backpropagation training process is the design of advanced optimisers to improve the convergence speed and stability of the gradient descent learning rule. In particular, in multi-task deep learning ($MTDL$) the multitude of tasks may generate potentially conflicting gradients that would hinder the concurrent convergence of the diverse loss functions. This challenge arises when the gradients of the task objectives have either different magnitudes or opposite directions, causing one or a few to dominate or to interfere with each other, thus degrading the training process. Gradient surgery methods address the problem explicitly dealing with conflicting gradients by adjusting the overall gradient trajectory. This work introduces a novel gradient surgery method, the Similarity-Aware Momentum Gradient Surgery (SAM-GS), which provides an effective and scalable approach based on a gradient magnitude similarity measure to guide the optimisation process. The SAM-GS surgery adopts gradient equalisation and modulation of the first-order momentum. A series of experimental tests have shown the effectiveness of SAM-GS on synthetic problems and $MTL$ benchmarks. Gradient magnitude similarity plays a crucial role in regularising gradient aggregation in $MTDL$ for the optimisation of the learning process. - 26 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Education
- Vol. 68 (2) , 248-257
Experimenting Team-Based Learning in a Large Computer Networks Class
Contribution: An innovative teaching experience carried out at the University of Trento using team-based learning (TBL) in a large computer networks class. The impact of TBL on the students’ learning and satisfaction was investigated. Background: Active learning pedagogies, including TBL, play an important role in enhancing higher-order cognitive skills among the student community. Reports on the implementation of TBL in engineering education are still scarce, despite its potential as an effective strategy for teaching problem solving skills in large classes. Intended Outcomes: Improved learning and student engagement via structured groupwork and challenging activities. A structure that makes it possible to scale TBL up to large classes with measurable learning improvements. Application Design: Laboratory classes were structured so that they became instrumental to TBL sessions, which in turn provided a stimulating environment to improve learning of key computer networks concepts. Grouping students with different backgrounds and previous knowledge enabled a more effective group work. Findings: The application of TBL fostered a deeper understanding of the topics covered in the course, resulting in higher scores on final exams and fewer failures. Interviewed students found the experience very satisfactory in terms of learning, group work, and involvement.
- 26 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Education
- Vol. 68 (2) , 248-257
Experimenting Team-Based Learning in a Large Computer Networks Class
Contribution: An innovative teaching experience carried out at the University of Trento using team-based learning (TBL) in a large computer networks class. The impact of TBL on the students’ learning and satisfaction was investigated. Background: Active learning pedagogies, including TBL, play an important role in enhancing higher-order cognitive skills among the student community. Reports on the implementation of TBL in engineering education are still scarce, despite its potential as an effective strategy for teaching problem solving skills in large classes. Intended Outcomes: Improved learning and student engagement via structured groupwork and challenging activities. A structure that makes it possible to scale TBL up to large classes with measurable learning improvements. Application Design: Laboratory classes were structured so that they became instrumental to TBL sessions, which in turn provided a stimulating environment to improve learning of key computer networks concepts. Grouping students with different backgrounds and previous knowledge enabled a more effective group work. Findings: The application of TBL fostered a deeper understanding of the topics covered in the course, resulting in higher scores on final exams and fewer failures. Interviewed students found the experience very satisfactory in terms of learning, group work, and involvement.
- Open Access
- 10 July 2006
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- No. 15582809,p. 157-162
Morphological Processing of Electromagnetic Scattering Data for Enhancing the Reconstruction Accuracy of the Iterative Multi-Scaling Approach
- 10 July 2006
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- No. 15582809,p. 157-162
Open AccessMorphological Processing of Electromagnetic Scattering Data for Enhancing the Reconstruction Accuracy of the Iterative Multi-Scaling Approach
- 1 June 2005
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-4
Improving the numerical effectiveness of a class of microwave imaging techniques for NDE/NDT with the SMW updating formula
This paper presents an innovative microwave imaging technique suitable for NDT/NDE applications. To reduce the number of problem unknowns as well a the computational load by improving the convergence rate of the process, the available a-priori information is exploited by introducing a computationally-effective procedure for the prediction of the electric field based on the updating Sherman-Morrison-Woodbury (SMW) formula. The numerical results as well as comparisons with state-of-the-art methods confirm the effectiveness, the feasibility, and the robustness of the proposed approach.
- 1 June 2005
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-4
Improving the numerical effectiveness of a class of microwave imaging techniques for NDE/NDT with the SMW updating formula
This paper presents an innovative microwave imaging technique suitable for NDT/NDE applications. To reduce the number of problem unknowns as well a the computational load by improving the convergence rate of the process, the available a-priori information is exploited by introducing a computationally-effective procedure for the prediction of the electric field based on the updating Sherman-Morrison-Woodbury (SMW) formula. The numerical results as well as comparisons with state-of-the-art methods confirm the effectiveness, the feasibility, and the robustness of the proposed approach.