Davide Anguita
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Davide Anguita has a total of 192 co-authors. Here are the top co-authors by number of shared publications:
- Luca Oneto
University of Genoa
76 shared publications - Sandro Ridella
DITEN Department, University of Genoa, Via Opera Pia 11A, I-16145, Italy
52 shared publications Alessandro Ghio
33 shared publicationsA. Boni
9 shared publications- Domenico Ponta
Università di Genova
9 shared publications - Giuliano Donzellini
University of Genova
9 shared publications
Latest Publications
- Open Access
- 1 November 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Intelligent Transportation Systems
- Vol. 27 (11) , 13704-13747
Trustworthy Data- and ML-Ops for Intelligent Transportation Systems and Logistics
- 1 November 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Intelligent Transportation Systems
- Vol. 27 (11) , 13704-13747
Open AccessTrustworthy Data- and ML-Ops for Intelligent Transportation Systems and Logistics
- 15 September 2026
- Published by Springer Nature
- p. 506-518
Instance Reweighting for Adversarial Training Helps Only Against the Training Attack
- 15 September 2026
- Published by Springer Nature
- p. 506-518
Instance Reweighting for Adversarial Training Helps Only Against the Training Attack
- Open Access
- 6 September 2026
- Published by Springer Nature in Complex & Intelligent Systems
Fair backward compatibility: definitions, theoretical framework, and empirical results
- 6 September 2026
- Published by Springer Nature in Complex & Intelligent Systems
Open AccessFair backward compatibility: definitions, theoretical framework, and empirical results
- Open Access
- 28 January 2026
- Published by Elsevier in Neurocomputing
- Vol. 674
Reconciling grokking with statistical learning theory through the lens of norm- and stability-based generalization bounds
- 28 January 2026
- Published by Elsevier in Neurocomputing
- Vol. 674
Open AccessReconciling grokking with statistical learning theory through the lens of norm- and stability-based generalization bounds
- Open Access
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 99841-99860
Overvalidation in Machine Learning: Empirical Evidence and Theoretical Insights
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 99841-99860
Open AccessOvervalidation in Machine Learning: Empirical Evidence and Theoretical Insights
- 1 October 2025
- Published by Springer Nature
- p. 379-391
Physics Informed Machine Learning for Power Flow Analysis: Injecting Knowledge via Pre-, In-, and Post-processing
- 1 October 2025
- Published by Springer Nature
- p. 379-391
Physics Informed Machine Learning for Power Flow Analysis: Injecting Knowledge via Pre-, In-, and Post-processing
- 23 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 47 (9) , 7457-7469
Robustness-Congruent Adversarial Training for Secure Machine Learning Model Updates
Machine-learning models demand periodic updates to improve their average accuracy, exploiting novel architectures and additional data. However, a newly updated model may commit mistakes the previous model did not make. Such misclassifications are referred to as negative flips, experienced by users as a regression of performance. In this work, we show that this problem also affects robustness to adversarial examples, hindering the development of secure model update practices. In particular, when updating a model to improve its adversarial robustness, previously ineffective adversarial attacks on some inputs may become successful, causing a regression in the perceived security of the system. We propose a novel technique, named robustness-congruent adversarial training, to address this issue. It amounts to fine-tuning a model with adversarial training, while constraining it to retain higher robustness on the samples for which no adversarial example was found before the update. We show that our algorithm and, more generally, learning with non-regression constraints, provides a theoretically-grounded framework to train consistent estimators. Our experiments on robust models for computer vision confirm that both accuracy and robustness, even if improved after model update, can be affected by negative flips, and our robustness-congruent adversarial training can mitigate the problem, outperforming competing baseline methods.
- 23 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 47 (9) , 7457-7469
Robustness-Congruent Adversarial Training for Secure Machine Learning Model Updates
Machine-learning models demand periodic updates to improve their average accuracy, exploiting novel architectures and additional data. However, a newly updated model may commit mistakes the previous model did not make. Such misclassifications are referred to as negative flips, experienced by users as a regression of performance. In this work, we show that this problem also affects robustness to adversarial examples, hindering the development of secure model update practices. In particular, when updating a model to improve its adversarial robustness, previously ineffective adversarial attacks on some inputs may become successful, causing a regression in the perceived security of the system. We propose a novel technique, named robustness-congruent adversarial training, to address this issue. It amounts to fine-tuning a model with adversarial training, while constraining it to retain higher robustness on the samples for which no adversarial example was found before the update. We show that our algorithm and, more generally, learning with non-regression constraints, provides a theoretically-grounded framework to train consistent estimators. Our experiments on robust models for computer vision confirm that both accuracy and robustness, even if improved after model update, can be affected by negative flips, and our robustness-congruent adversarial training can mitigate the problem, outperforming competing baseline methods.
- Open Access
- 7 May 2024
- Published by Elsevier in Neurocomputing
- Vol. 592
Towards algorithms and models that we can trust: A theoretical perspective
- 7 May 2024
- Published by Elsevier in Neurocomputing
- Vol. 592
Open AccessTowards algorithms and models that we can trust: A theoretical perspective
- 31 July 2023
- Published by Springer Nature in Neural Processing Letters
- Vol. 55 (5) , 5269-5272
New Advances in Artificial Neural Networks and Machine Learning Techniques
- 31 July 2023
- Published by Springer Nature in Neural Processing Letters
- Vol. 55 (5) , 5269-5272
New Advances in Artificial Neural Networks and Machine Learning Techniques