Four IRT AESE – Saint Exupéry and ANITI scientific publications, in the framework of the DEEL project have been selected at this year’s NeurIPS conference:
- Numerical influence of ReLU'(0) on backpropagation. NeurIPS 2021, David Bertoin, Sébastien Gerchinovitz, Jérôme Bolte, Edouard Pauwels.
- Look at the Variance! Efficient Black-box Explanations with Sobol-based Sensitivity Analysis. NeurIPS 2021, David Vigouroux, Thomas FEL, Thomas Serre, Remi Cadene, Mathieu Chalvidal, Matthieu Cord
- Instance-Dependent Bounds for Zeroth-order Lipschitz Optimization with Error Certificates. NeurIPS 2021, Sébastien Gerchinovitz, François Bachoc, Tommaso Cesari.
- Semialgebraic Representation of Monotone Deep Equilibrium Models and Applications to Certification. NeurIPS 2021, Tong Chen, Jean-Bernard Lasserre, Victor Magron, Edouard Pauwels.
- Nonsmooth Implicit Differentiation for Machine Learning and Optimization. Jérôme Bolte, Tâm Lê, Edouard Pauwels and Antonio Silveti-Falls.
Congratulations to all for these great achievement and contributions to the field of Artificial Intelligence research.
English 
French