Training Graph Neural Networks Subject to a Tight Lipschitz Constraint

S Juvina, AA Neacșu, JC Pesquet… - … on Machine Learning …, 2024 - inria.hal.science
We propose a strategy for training a wide range of graph neural networks (GNNs) under tight
Lipschitz bound constraints. Specifically, by leveraging graph spectral theory, we derive …

A quantitative analysis of the robustness of neural networks for tabular data

K Gupta, B Pesquet-Popescu, F Kaakai… - ICASSP 2021-2021 …, 2021 - ieeexplore.ieee.org
This paper presents a quantitative approach to demonstrate the robustness of neural
networks for tabular data. These data form the backbone of the data structures found in most …

Design of Robust Complex-Valued Feed-Forward Neural Networks

A Neacşu, R Ciubotaru, JC Pesquet… - 2022 30th European …, 2022 - ieeexplore.ieee.org
This paper addresses the problem of designing robust complex-valued neural networks in
order to reduce their sensitivity to adversarial perturbations. The robustness is guaranteed …

Testing the Robustness of Deepfake Detectors

A Radu, A Neacşu - 2024 15th International Conference on …, 2024 - ieeexplore.ieee.org
The term deepfake is used to denote an artificially generated or altered image using deep
neural networks. Such methods are widely spread, with a focus on creating more realistic …

Leveraging end-to-end denoisers for denoising periodic signals

J Rio, O Alata, F Momey… - 2021 29th European Signal …, 2021 - ieeexplore.ieee.org
In this paper, we propose a new framework for denoising 1D periodic signals with deep
learning models by exploiting their periodic properties. Our method lies on a transformation …

Stability Quantification of Neural Networks

K Gupta - 2023 - theses.hal.science
Artificial neural networks are at the core of recent advances in Artificial Intelligence. One of
the main challenges faced today, especially by companies likeThales designing advanced …