Massimiliano Mancini

Unlearning Under Imbalance: Benchmarking Fairness in Multimodal LLM Unlearning

Lorenzo Orsingher, Thomas De Min, Massimiliano Mancini, et al.

Harnessing Textual Refusal Directions for Multimodal Safety

Moreno D'Incà, Massimiliano Mancini, Nicu Sebe

Structural Pruning of Large Vision Language Models: A Comprehensive Study on Pruning Dynamics, Recovery, and Data Efficiency

Yiran Huang, Lukas Thede, Massimiliano Mancini, et al.

SEM: Sparse Embedding Modulation for Post-Hoc Debiasing of Vision-Language Models

Quentin Guimard, Federico Bartsch, Simone Caldarella, et al.

This paper introduces SEM (Sparse Embedding Modulation), a method to remove social and spurious biases from vision-language models like CLIP without hurting their performance. Instead of working directly with CLIP's dense embeddings where bias is tangled with useful information, SEM uses a Sparse Autoencoder to break down embeddings into cleaner, separate features, allowing it to precisely remove bias-related features while keeping task-relevant ones intact. The method shows strong improvements in fairness across multiple benchmarks without requiring retraining.