AIConfPaper
← Search

Pau de Jorge

5 accepted papers

2025

DUNE: Distilling a Universal Encoder from Heterogeneous 2D and 3D Teachers

Mert Bülent Sarıyıldız, Philippe Weinzaepfel, Thomas Lucas, Pau de Jorge, Diane Larlus, Yannis Kalantidis
CVPR 2025poster

Recent multi-teacher distillation methods have unified the encoders of multiple foundation models into a single encoder, achieving competitive performance on core vision tasks like classification, segmentation, and depth estimation. This led us to ask: Could similar success be achieved when the pool…

Cited by 0Source ↗PDF ↗Scholar ↗
2023

Reliability in Semantic Segmentation: Are We on the Right Track?

Pau de Jorge, Riccardo Volpi, Philip H.S. Torr, Grégory Rogez
CVPR 2023poster

Motivated by the increasing popularity of transformers in computer vision, in recent times there has been a rapid development of novel architectures. While in-domain performance follows a constant, upward trend, properties like robustness or uncertainty estimation are less explored -leaving doubts a…

Cited by 29Source ↗PDF ↗code ↗Scholar ↗
2022

Make Some Noise: Reliable and Efficient Single-Step Adversarial Training

Pau de Jorge, Adel Bibi, Riccardo Volpi, Amartya Sanyal, Philip Torr, Grégory Rogez +1
NeurIPS 2022accept

Recently, Wong et al. (2020) showed that adversarial training with single-step FGSM leads to a characteristic failure mode named catastrophic overfitting (CO), in which a model becomes suddenly vulnerable to multi-step attacks. Experimentally they showed that simply adding a random perturbation prio…

Cited by 63Source ↗PDF ↗code ↗Scholar ↗
2022

On the Road to Online Adaptation for Semantic Image Segmentation

Riccardo Volpi, Pau De Jorge, Diane Larlus, Gabriela Csurka
CVPR 2022poster

We propose a new problem formulation and a corresponding evaluation framework to advance research on unsupervised domain adaptation for semantic image segmentation. The overall goal is fostering the development of adaptive learning systems that will continuously learn, without supervision, in ever-c…

Cited by 35PDF ↗code ↗Scholar ↗
2021

Progressive Skeletonization: Trimming more fat from a network at initialization

Pau de Jorge, Amartya Sanyal, Harkirat Behl, Philip Torr, Grégory Rogez, Puneet K. Dokania
ICLR 2021poster

Recent studies have shown that skeletonization (pruning parameters) of networks at initialization provides all the practical benefits of sparsity both at inference and training time, while only marginally degrading their performance. However, we observe that beyond a certain level of sparsity (appro…

Cited by 118Source ↗PDF ↗code ↗Scholar ↗