Reliability

When the unexpected happens, when the weather badly degrades, when a sensor gets blocked, the embarked perception system should diagnose the situation and react accordingly, e.g., by calling an alternative system or the human driver. With this in mind, we investigate ways to improve the robustness of neural nets to input variations, including to adversarial attacks, and to predict automatically the performance and the confidence of their predictions as in ConfidNet at NeurIPS’19.

Publications

Latent Discriminant deterministic Uncertainty

Gianni Franchi, Andrei Bursuc, Emanuel Aldea, Severine Dubuisson, and David Filliat
European Conference on Computer Vision (ECCV), 2022

images/publications/2022_dtnu/dtnu.png publication teaser

Solving Disjunctive Temporal Networks with Uncertainty under Restricted Time-Based Controllability using Tree Search and Graph Neural Networks

Kevin Osanlou, Jeremy Frank, J. Benton, Andrei Bursuc, Christophe Guettier, Tristan Cazenave and Eric Jacopin
AAAI Conference on Artificial Intelligence (AAAI), 2022

images/publications/2021_superpixelmix/superpixelmix_teaser.jpg publication teaser

Reliability

Publications

Latent Discriminant deterministic Uncertainty

Solving Disjunctive Temporal Networks with Uncertainty under Restricted Time-Based Controllability using Tree Search and Graph Neural Networks

Robust Semantic Segmentation with Superpixel-Mix

Triggering Failures: Out-Of-Distribution detection by learning from local adversarial attacks in Semantic Segmentation

StyleLess layer: Improving robustness for real-world driving

Confidence Estimation via Auxiliary Models

TRADI: Tracking deep neural network weight distributions for uncertainty estimation

Addressing Failure Prediction by Learning Model Confidence

Optimal Solving of Constrained Path-Planning Problems with Graph Convolutional Networks and Optimized Tree Search