6 papers
Synesthesia via Direct Latent Augmentation:Bypassing the Decode-Encode Loop for Cross-Modal Distillation
Cristian Sbrolli, Nicolas Michel, Matteo Matteucci +1
While multimodal integration significantly improves computer vision models, deploying them incurs prohibitive inference costs and requires scarce, perfectly paired datasets. Recent…
Auto-Comp: An Automated Pipeline for Scalable Compositional Probing of Contrastive Vision-Language Models
Cristian Sbrolli, Matteo Matteucci, Toshihiko Yamasaki
Modern Vision-Language Models (VLMs) exhibit a critical flaw in compositional reasoning, often confusing "a red cube and a blue sphere" with "a blue cube and a red sphere". Disenta…
PolyGen: Fully Synthetic Vision-Language Training via Multi-Generator Ensembles
Leonardo Brusini, Cristian Sbrolli, Eugenio Lomurno +2
Synthetic data offers a scalable solution for vision-language pre-training, yet current state-of-the-art methods typically rely on scaling up a single generative backbone, which in…
SCENEFORGE: Enhancing 3D-text alignment with Structured Scene Compositions
Cristian Sbrolli, Matteo Matteucci
The whole is greater than the sum of its parts-even in 3D-text contrastive learning. We introduce SceneForge, a novel framework that enhances contrastive alignment between 3D point…
Your Image Generator Is Your New Private Dataset
Nicolo Resmini, Eugenio Lomurno, Cristian Sbrolli +1
Generative diffusion models have emerged as powerful tools to synthetically produce training data, offering potential solutions to data scarcity and reducing labelling costs for do…
Neuro-Symbolic Scene Graph Conditioning for Synthetic Image Dataset Generation
Giacomo Savazzi, Eugenio Lomurno, Cristian Sbrolli +2
As machine learning models increase in scale and complexity, obtaining sufficient training data has become a critical bottleneck due to acquisition costs, privacy constraints, and…