Lorenzo Bonicelli — accepted AI-conference papers

2026

Dataless Weight Disentanglement in Task Arithmetic via Kronecker-Factored Approximate Curvature

Angelo Porrello, Pietro Buzzega, Felix Dangel, Thomas Sommariva, Riccardo Salami, Lorenzo Bonicelli +1

ICLR 2026poster

Task Arithmetic yields a modular, scalable way to adapt foundation models. Combining multiple task vectors, however, can lead to cross-task interference, causing representation drift and degraded performance. Representation drift regularization provides a natural remedy to disentangle task vectors;…

Cited by 0 Source ↗Scholar ↗

2025

A Second-Order Perspective on Model Compositionality and Incremental Learning

Angelo Porrello, Lorenzo Bonicelli, Pietro Buzzega, Monica Millunzi, Simone Calderara, Rita Cucchiara

ICLR 2025spotlight

The fine-tuning of deep pre-trained models has revealed compositional properties, with multiple specialized modules that can be arbitrarily composed into a single, multi-task model. However, identifying the conditions that promote compositionality remains an open issue, with recent efforts concentra…

Cited by 4 Source ↗PDF ↗code ↗Scholar ↗

2025

Update Your Transformer to the Latest Release: Re-Basin of Task Vectors

Filippo Rinaldi, Giacomo Capitani, Lorenzo Bonicelli, Donato Crisostomi, Federico Bolelli, ELISA FICARRA +3

ICML 2025poster

Foundation models serve as the backbone for numerous specialized models developed through fine-tuning. However, when the underlying pretrained model is updated or retrained (e.g., on larger and more curated datasets), the fine-tuned model becomes obsolete, losing its utility and requiring retraining…

Cited by 0 Source ↗PDF ↗code ↗Scholar ↗

2024

Saliency-driven Experience Replay for Continual Learning

Giovanni Bellitto, Federica Proietto Salanitri, Matteo Pennisi, Matteo Boschini, Lorenzo Bonicelli, Angelo Porrello +3

NeurIPS 2024spotlight

We present Saliency-driven Experience Replay - SER - a biologically-plausible approach based on replicating human visual saliency to enhance classification models in continual learning settings. Inspired by neurophysiological evidence that the primary visual cortex does not contribute to object mani…

Cited by 3 Source ↗PDF ↗code ↗Scholar ↗

2024

Semantic Residual Prompts for Continual Learning

Martin Menabue*, Emanuele Frascaroli, Matteo Boschini, Enver Sangineto, Lorenzo Bonicelli, Angelo Porrello* +1

ECCV 2024poster

"Prompt-tuning methods for Continual Learning (CL) freeze a large pre-trained model and train a few parameter vectors termed prompts. Most of these methods organize these vectors in a pool of key-value pairs and use the input image as query to retrieve the prompts (values). However, as keys are lear…

Cited by 11 Source ↗PDF ↗code ↗Scholar ↗

2022

On the Effectiveness of Lipschitz-Driven Rehearsal in Continual Learning

Lorenzo Bonicelli, Matteo Boschini, Angelo Porrello, Concetto Spampinato, Simone Calderara

NeurIPS 2022accept

Rehearsal approaches enjoy immense popularity with Continual Learning (CL) practitioners. These methods collect samples from previously encountered data distributions in a small memory buffer; subsequently, they repeatedly optimize on the latter to prevent catastrophic forgetting. This work draws at…

Cited by 55 Source ↗PDF ↗code ↗Scholar ↗

2022

Transfer without Forgetting

Matteo Boschini, Lorenzo Bonicelli, Angelo Porrello, Giovanni Bellitto, Matteo Pennisi, Simone Palazzo +2

ECCV 2022poster

"This work investigates the entanglement between Continual Learning (CL) and Transfer Learning (TL). In particular, we shed light on the widespread application of network pretraining, highlighting that it is itself subject to catastrophic forgetting. Unfortunately, this issue leads to the under-expl…

Cited by 57 Source ↗PDF ↗code ↗Scholar ↗